We describe our effort on automated extraction of socio-political events from news in the scope of a workshop and a shared task we organized at Language Resources and Evaluation Conference (LREC 2020). We believe the event extraction studies in computational linguistics and social and political sciences should further support each other in order to enable large scale socio-political event information collection across sources, countries, and languages. The event consists of regular research papers and a shared task, which is about event sentence coreference identification (ESCI), tracks. All submissions were reviewed by five members of the program committee. The workshop attracted research papers related to evaluation of machine learning methodologies, language resources, material conflict forecasting, and a shared task participation report in the scope of socio-political event information collection. It has shown us the volume and variety of both the data sources and event information collection approaches related to socio-political events and the need to fill the gap between automated text processing techniques and requirements of social and political sciences.
Not all conflict datasets offer equal levels of coverage, depth, use-ability, and content. A review of the inclusion criteria, methodology, and sourcing of leading publicly available conflict datasets demonstrates that there are significant discrepancies in the output produced by ostensibly similar projects. This keynote will question the presumption of substantial overlap between datasets, and identify a number of important gaps left by deficiencies across core criteria for effective conflict data collection and analysis.
In this brief keynote, I will address what I see as five majorissues in terms of development for operational event datasets (that is, event data intended for real time monitoringand forecasting, rather than purely for academic research).
This study evaluates the robustness of two state-of-the-art deep contextual language representations, ELMo and DistilBERT, on supervised learning of binary protest news classification (PC) and sentiment analysis (SA) of product reviews. A ”cross-context” setting is enabled using test sets that are distinct from the training data. The models are fine-tuned and fed into a Feed-Forward Neural Network (FFNN) and a Bidirectional Long Short Term Memory network (BiLSTM). Multinomial Naive Bayes (MNB) and Linear Support Vector Machine (LSVM) are used as traditional baselines. The results suggest that DistilBERT can transfer generic semantic knowledge to other domains better than ELMo. DistilBERT is also 30% smaller and 83% faster than ELMo, which suggests superiority for smaller computational training budgets. When generalization is not the utmost preference and test domain is similar to the training domain, the traditional machine learning (ML) algorithms can still be considered as more economic alternatives to deep language representations.
We cast the problem of event annotation as one of text categorization, and compare state of the art text categorization techniques on event data produced within the Uppsala Conflict Data Program (UCDP). Annotating a single text involves assigning the labels pertaining to at least 17 distinct categorization tasks, e.g., who were the attacking organization, who was attacked, and where did the event take place. The text categorization techniques under scrutiny are a classical Bag-of-Words approach; character-based contextualized embeddings produced by ELMo; embeddings produced by the BERT base model, and a version of BERT base fine-tuned on UCDP data; and a pre-trained and fine-tuned classifier based on ULMFiT. The categorization tasks are very diverse in terms of the number of classes to predict as well as the skeweness of the distribution of classes. The categorization results exhibit a large variability across tasks, ranging from 30.3% to 99.8% F-score.
Automating the detection of event mentions in online texts and their classification vis-a-vis domain-specific event type taxonomies has been acknowledged by many organisations worldwide to be of paramount importance in order to facilitate the process of intelligence gathering. This paper reports on some preliminary experiments of comparing various linguistically-lightweight approaches for fine-grained event classification based on short text snippets reporting on events. In particular, we compare the performance of a TF-IDF-weighted character n-gram SVM-based model versus SVMs trained on various of-the-shelf pre-trained word embeddings (GloVe, BERT, FastText) as features. We exploit a relatively large event corpus consisting of circa 610K short text event descriptions classified using a 25-event categories that cover political violence and protest events. The best results, i.e., 83.5% macro and 92.4% micro F1 score, were obtained using the TF-IDF-weighted character n-gram model.
Previous efforts to automate the detection of social and political events in text have primarily focused on identifying events described within single sentences or documents. Within a corpus of documents, these automated systems are unable to link event references—recognize singular events across multiple sentences or documents. A separate literature in computational linguistics on event coreference resolution attempts to link known events to one another within (and across) documents. I provide a data set for evaluating methods to identify certain political events in text and to link related texts to one another based on shared events. The data set, Headlines of War, is built on the Militarized Interstate Disputes data set and offers headlines classified by dispute status and headline pairs labeled with coreference indicators. Additionally, I introduce a model capable of accomplishing both tasks. The multi-task convolutional neural network is shown to be capable of recognizing events and event coreferences given the headlines’ texts and publication dates.
This paper presents the conflict event modelling experiment, conducted at the Joint Research Centre of the European Commission, particularly focusing on the limitations of the input data. This model is under evaluation as to potentially complement the Global Conflict Risk Index (GCRI), a conflict risk model supporting the design of European Union’s conflict prevention strategies. The model aims at estimating the occurrence of material conflict events, under the assumption that an increase in material conflict events goes along with a decrease in material and verbal cooperation. It adopts a Long-Short Term Memory Cell Recurrent Neural Network on country-level actor-based event datasets that indicate potential triggers to violent conflict such as demonstrations, strikes, or elections-related violence. The observed data and the outcome of the model predictions consecutively, consolidate an early warning alarm system that signals abnormal social unrest upheavals, and appears promising as an approach towards a conflict trigger model. However, event-based systems still require overcoming certain obstacles related to the quality of the input data and the event classification method.
This article introduces Hadath, a supervised protocol for coding event data from text written in Arabic. Hadath contributes to recent efforts in advancing multi-language event coding using computer-based solutions. In this application, we focus on extracting event data about the conflict in Afghanistan from 2008 to 2018 using Arabic information sources. The implementation relies first on a Machine Learning algorithm to classify news stories relevant to the Afghan conflict. Then, using Hadath, we implement the Natural Language Processing component for event coding from Arabic script. The output database contains daily geo-referenced information at the district level on who did what to whom, when and where in the Afghan conflict. The data helps to identify trends in the dynamics of violence, the provision of governance, and traditional conflict resolution in Afghanistan for different actors over time and across space.
The advent of Big Data has shifted social science research towards computational methods. The volume of data that is nowadays available has brought a radical change in traditional approaches due to the cost and effort needed for processing. Knowledge extraction from heterogeneous and ample data is not an easy task to tackle. Thus, interdisciplinary approaches are necessary, combining experts of both social and computer science. This paper aims to present a work in the context of protest analysis, which falls into the scope of Computational Social Science. More specifically, the contribution of this work is to describe a Computational Social Science methodology for Event Analysis. The presented methodology is generic in the sense that it can be applied in every event typology and moreover, it is innovative and suitable for interdisciplinary tasks as it incorporates the human-in-the-loop. Additionally, a case study is presented concerning Protest Analysis in Greece over the last two decades. The conceptual foundation lies mainly upon claims analysis, and newspaper data were used in order to map, document and discuss protests in Greece in a longitudinal perspective.
This paper summarizes our group’s efforts in the event sentence coreference identification shared task, which is organized as part of the Automated Extraction of Socio-Political Events from News (AESPEN) Workshop. Our main approach consists of three steps. We initially use a transformer based model to predict whether a pair of sentences refer to the same event or not. Later, we use these predictions as the initial scores and recalculate the pair scores by considering the relation of sentences in a pair with respect to other sentences. As the last step, final scores between these sentences are used to construct the clusters, starting with the pairs with the highest scores. Our proposed approach outperforms the baseline approach across all evaluation metrics.
Cultural institutions such as galleries, libraries, archives and museums continue to make commitments to large scale digitization of collections. An ongoing challenge is how to increase discovery and access through structured data and the semantic web. In this paper we describe a method for using computer vision algorithms that automatically detect regions of “stuff” — such as the sky, water, and roads — to produce rich and accurate structured data triples for describing the content of historic photography. We apply our method to a collection of 1610 documentary photographs produced in the 1930s and 1940 by the FSA-OWI division of the U.S. federal government. Manual verification of the extracted annotations yields an accuracy rate of 97.5%, compared to 70.7% for relations extracted from object detection and 31.5% for automatically generated captions. Our method also produces a rich set of features, providing more unique labels (1170) than either the captions (1040) or object detection (178) methods. We conclude by describing directions for a linguistically-focused ontology of region categories that can better enrich historical image data. Open source code and the extracted metadata from our corpus are made available as external resources.
Iconclass, being a a well established classification system, could benefit from interconnections with other ontologies in order to semantically enrich its content. This work presents a disambiguating and interlinking approach which is used to map Iconclass Subjects to concepts of the Art and Architecture Thesaurus. In a preliminary evaluation, the system is able to produce promising predictions, though the task is highly challenging due to conceptual and schema heterogeneity. Several algorithmic improvements for this specific interlinking task, as well as and future research directions are suggestions. The produced mappings, as well as the source code and additional information can be found at https://github.com/annabreit/taxonomy-interlinking.
The aim of this position paper is to establish an initial approach to the automatic classification of digital images about the Outsider Art style of painting. Specifically, we explore whether is it possible to classify non-traditional artistic styles by using the same features that are used for classifying traditional styles? Our research question is motivated by two facts. First, art historians state that non-traditional styles are influenced by factors “outside” of the world of art. Second, some studies have shown that several artistic styles confound certain classification techniques. Following current approaches to style prediction, this paper utilises Deep Learning methods to encode image features. Our preliminary experiments have provided motivation to think that, as is the case with traditional styles, Outsider Art can be computationally modelled with objective means by using training datasets and CNN models. Nevertheless, our results are not conclusive due to the lack of a large available dataset on Outsider Art. Therefore, at the end of the paper, we have mapped future lines of action, which include the compilation of a large dataset of Outsider Art images and the creation of an ontology of Outsider Art.
Cultural heritage data plays a pivotal role in the understanding of human history and culture. A wealth of information is buried in art-historic archives which can be extracted via digitization and analysis. This information can facilitate search and browsing, help art historians to track the provenance of artworks and enable wider semantic text exploration for digital cultural resources. However, this information is contained in images of artworks, as well as textual descriptions or annotations accompanied with the images. During the digitization of such resources, the valuable associations between the images and texts are frequently lost. In this project description, we propose an approach to retrieve the associations between images and texts for artworks from art-historic archives. To this end, we use machine learning to generate text descriptions for the extracted images on the one hand, and to detect descriptive phrases and titles of images from the text on the other hand. Finally, we use embeddings to align both, the descriptions and the images.
Semantic enrichment of historical images to build interactive AI systems for the Digital Humanities domain has recently gained significant attention. However, before implementing any semantic enrichment tool for building AI systems, it is also crucial to analyse the quality and richness of the existing datasets and understand the areas where semantic enrichment is most required. Here, we propose an approach to conducting a preliminary analysis of selected historical images from the Europeana platform using existing linked data quality assessment tools. The analysis targets food images by collecting metadata provided from curators such as Galleries, Libraries, Archives and Museums (GLAMs) and cultural aggregators such as Europeana. We identified metrics to evaluate the quality of the metadata associated with food-related images which are harvested from the Europeana platform. In this paper, we present the food-image dataset, the associated metadata and our proposed method for the assessment. The results of our assessment will be used to guide the current effort to semantically enrich the images and build high-quality metadata using Computer Vision.
ImageNet has millions of images that are labeled with English WordNet synsets. This paper investigates the extension of ImageNet to Arabic using Arabic WordNet. The objective is to discover if Arabic synsets can be found for synsets used in ImageNet. The primary finding is the identification of Arabic synsets for 1,219 of the 21,841 synsets used in ImageNet, which represents 1.1 million images. By leveraging the parent-child structure of synsets in ImageNet, this dataset is extended to 10,462 synsets (and 7.1 million images) that have an Arabic label, which is either a match or a direct hypernym, and to 17,438 synsets (and 11 million images) when a hypernym of a hypernym is included. When all hypernyms for a node are considered, an Arabic synset is found for all but four synsets. This represents the major contribution of this work: a dataset of images that have Arabic labels for 99.9% of the images in ImageNet.
Scene graph is a graph representation that explicitly represents high-level semantic knowledge of an image such as objects, attributes of objects and relationships between objects. Various tasks have been proposed for the scene graph, but the problem is that they have a limited vocabulary and biased information due to their own hypothesis. Therefore, results of each task are not generalizable and difficult to be applied to other down-stream tasks. In this paper, we propose Entity Synset Alignment(ESA), which is a method to create a general scene graph by aligning various semantic knowledge efficiently to solve this bias problem. The ESA uses a large-scale lexical database, WordNet and Intersection of Union (IoU) to align the object labels in multiple scene graphs/semantic knowledge. In experiment, the integrated scene graph is applied to the image-caption retrieval task as a down-stream task. We confirm that integrating multiple scene graphs helps to get better representations of images.
Visual Question Generation (VQG), the task of generating a question based on image contents, is an increasingly important area that combines natural language processing and computer vision. Although there are some recent works that have attempted to generate questions from images in the open domain, the task of VQG in the medical domain has not been explored so far. In this paper, we introduce an approach to generation of visual questions about radiology images called VQGR, i.e. an algorithm that is able to ask a question when shown an image. VQGR first generates new training data from the existing examples, based on contextual word embeddings and image augmentation techniques. It then uses the variational auto-encoders model to encode images into a latent space and decode natural language questions. Experimental automatic evaluations performed on the VQA-RAD dataset of clinical visual questions show that VQGR achieves good performances compared with the baseline system. The source code is available at https://github.com/sarrouti/vqgr.
Task success is the standard metric used to evaluate referential visual dialogue systems. In this paper we propose two new metrics that evaluate how each question contributes to the goal. First, we measure how effective each question is by evaluating whether the question discards objects that are not the referent. Second, we define referring questions as those that univocally identify one object in the image. We report the new metrics for human dialogues and for state of the art publicly available models on GuessWhat?!. Regarding our first metric, we find that successful dialogues do not have a higher percentage of effective questions for most models. With respect to the second metric, humans make questions at the end of the dialogue that are referring, confirming their guess before guessing. Human dialogues that use this strategy have a higher task success but models do not seem to learn it.
We propose a novel alignment mechanism to deal with procedural reasoning on a newly released multimodal QA dataset, named RecipeQA. Our model is solving the textual cloze task which is a reading comprehension on a recipe containing images and instructions. We exploit the power of attention networks, cross-modal representations, and a latent alignment space between instructions and candidate answers to solve the problem. We introduce constrained max-pooling which refines the max pooling operation on the alignment matrix to impose disjoint constraints among the outputs of the model. Our evaluation result indicates a 19% improvement over the baselines.
Simultaneous Translation is a great challenge in which translation starts before the source sentence finished. Most studies take transcription as input and focus on balancing translation quality and latency for each sentence. However, most ASR systems can not provide accurate sentence boundaries in realtime. Thus it is a key problem to segment sentences for the word streaming before translation. In this paper, we propose a novel method for sentence boundary detection that takes it as a multi-class classification task under the end-to-end pre-training framework. Experiments show significant improvements both in terms of translation quality and latency.
End-to-End speech translation usually leverages audio-to-text parallel data to train an available speech translation model which has shown impressive results on various speech translation tasks. Due to the artificial cost of collecting audio-to-text parallel data, the speech translation is a natural low-resource translation scenario, which greatly hinders its improvement. In this paper, we proposed a new adversarial training method to leverage target monolingual data to relieve the low-resource shortcoming of speech translation. In our method, the existing speech translation model is considered as a Generator to gain a target language output, and another neural Discriminator is used to guide the distinction between outputs of speech translation model and true target monolingual sentences. Experimental results on the CCMT 2019-BSTC dataset speech translation task demonstrate that the proposed methods can significantly improve the performance of the End-to-End speech translation system.
In many practical applications, neural machine translation systems have to deal with the input from automatic speech recognition (ASR) systems which may contain a certain number of errors. This leads to two problems which degrade translation performance. One is the discrepancy between the training and testing data and the other is the translation error caused by the input errors may ruin the whole translation. In this paper, we propose a method to handle the two problems so as to generate robust translation to ASR errors. First, we simulate ASR errors in the training data so that the data distribution in the training and test is consistent. Second, we focus on ASR errors on homophone words and words with similar pronunciation and make use of their pronunciation information to help the translation model to recover from the input errors. Experiments on two Chinese-English data sets show that our method is more robust to input errors and can outperform the strong Transformer baseline significantly.
Autoregressive neural machine translation (NMT) models are often used to teach non-autoregressive models via knowledge distillation. However, there are few studies on improving the quality of autoregressive translation (AT) using non-autoregressive translation (NAT). In this work, we propose a novel Encoder-NAD-AD framework for NMT, aiming at boosting AT with global information produced by NAT model. Specifically, under the semantic guidance of source-side context captured by the encoder, the non-autoregressive decoder (NAD) first learns to generate target-side hidden state sequence in parallel. Then the autoregressive decoder (AD) performs translation from left to right, conditioned on source-side and target-side hidden states. Since AD has global information generated by low-latency NAD, it is more likely to produce a better translation with less time delay. Experiments on WMT14 En-De, WMT16 En-Ro, and IWSLT14 De-En translation tasks demonstrate that our framework achieves significant improvements with only 8% speed degeneration over the autoregressive NMT.
Recently, document-level neural machine translation (NMT) has become a hot topic in the community of machine translation. Despite its success, most of existing studies ignored the discourse structure information of the input document to be translated, which has shown effective in other tasks. In this paper, we propose to improve document-level NMT with the aid of discourse structure information. Our encoder is based on a hierarchical attention network (HAN) (Miculicich et al., 2018). Specifically, we first parse the input document to obtain its discourse structure. Then, we introduce a Transformer-based path encoder to embed the discourse structure information of each word. Finally, we combine the discourse structure information with the word embedding before it is fed into the encoder. Experimental results on the English-to-German dataset show that our model can significantly outperform both Transformer and Transformer+HAN.
This paper describes our machine translation systems for the streaming Chinese-to-English translation task of AutoSimTrans 2020. We present a sentence length based method and a sentence boundary detection model based method for the streaming input segmentation. Experimental results of the transcription and the ASR output translation on the development data sets show that the translation system with the detection model based method outperforms the one with the length based method in BLEU score by 1.19 and 0.99 respectively under similar or better latency.
Readability assessment aims to automatically classify text by the level appropriate for learning readers. Traditional approaches to this task utilize a variety of linguistically motivated features paired with simple machine learning models. More recent methods have improved performance by discarding these features and utilizing deep learning models. However, it is unknown whether augmenting deep learning models with linguistically motivated features would improve performance further. This paper combines these two approaches with the goal of improving overall model performance and addressing this question. Evaluating on two large readability corpora, we find that, given sufficient training data, augmenting deep learning models with linguistically motivated features does not improve state-of-the-art performance. Our results provide preliminary evidence for the hypothesis that the state-of-the-art deep learning models represent linguistic features of the text related to readability. Future research on the nature of representations formed in these models can shed light on the learned features and their relations to linguistically motivated ones hypothesized in traditional approaches.
The effect of noisy labels on the performance of NLP systems has been studied extensively for system training. In this paper, we focus on the effect that noisy labels have on system evaluation. Using automated scoring as an example, we demonstrate that the quality of human ratings used for system evaluation have a substantial impact on traditional performance metrics, making it impossible to compare system evaluations on labels with different quality. We propose that a new metric, PRMSE, developed within the educational measurement community, can help address this issue, and provide practical guidelines on using PRMSE.
Automated Essay Scoring (AES) can be used to automatically generate holistic scores with reliability comparable to human scoring. In addition, AES systems can provide formative feedback to learners, typically at the essay level. In contrast, we are interested in providing feedback specialized to the content of the essay, and specifically for the content areas required by the rubric. A key objective is that the feedback should be localized alongside the relevant essay text. An important step in this process is determining where in the essay the rubric designated points and topics are discussed. A natural approach to this task is to train a classifier using manually annotated data; however, collecting such data is extremely resource intensive. Instead, we propose a method to predict these annotation spans without requiring any labeled annotation data. Our approach is to consider AES as a Multiple Instance Learning (MIL) task. We show that such models can both predict content scores and localize content by leveraging their sentence-level score predictions. This capability arises despite never having access to annotation training data. Implications are discussed for improving formative feedback and explainable AES models.
Increased demand to learn English for business and education has led to growing interest in automatic spoken language assessment and teaching systems. With this shift to automated approaches it is important that systems reliably assess all aspects of a candidate’s responses. This paper examines one form of spoken language assessment; whether the response from the candidate is relevant to the prompt provided. This will be referred to as off-topic spoken response detection. Two forms of previously proposed approaches are examined in this work: the hierarchical attention-based topic model (HATM); and the similarity grid model (SGM). The work focuses on the scenario when the prompt, and associated responses, have not been seen in the training data, enabling the system to be applied to new test scripts without the need to collect data or retrain the model. To improve the performance of the systems for unseen prompts, data augmentation based on easy data augmentation (EDA) and translation based approaches are applied. Additionally for the HATM, a form of prompt dropout is described. The systems were evaluated on both seen and unseen prompts from Linguaskill Business and General English tests. For unseen data the performance of the HATM was improved using data augmentation, in contrast to the SGM where no gains were obtained. The two approaches were found to be complementary to one another, yielding a combined F0.5 score of 0.814 for off-topic response detection where the prompts have not been seen in training.
One-to-one tutoring is often an effective means to help students learn, and recent experiments with neural conversation systems are promising. However, large open datasets of tutoring conversations are lacking. To remedy this, we propose a novel asynchronous method for collecting tutoring dialogue via crowdworkers that is both amenable to the needs of deep learning algorithms and reflective of pedagogical concerns. In this approach, extended conversations are obtained between crowdworkers role-playing as both students and tutors. The CIMA collection, which we make publicly available, is novel in that students are exposed to overlapping grounded concepts between exercises and multiple relevant tutoring responses are collected for the same input. CIMA contains several compelling properties from an educational perspective: student role-players complete exercises in fewer turns during the course of the conversation and tutor players adopt strategies that conform with some educational conversational norms, such as providing hints versus asking questions in appropriate contexts. The dataset enables a model to be trained to generate the next tutoring utterance in a conversation, conditioned on a provided action strategy.
In this paper we employ a novel approach to advancing our understanding of the development of writing in English and German children across school grades using classification tasks. The data used come from two recently compiled corpora: The English data come from the the GiC corpus (983 school children in second-, sixth-, ninth- and eleventh-grade) and the German data are from the FD-LEX corpus (930 school children in fifth- and ninth-grade). The key to this paper is the combined use of what we refer to as ‘complexity contours’, i.e. series of measurements that capture the progression of linguistic complexity within a text, and Recurrent Neural Network (RNN) classifiers that adequately capture the sequential information in those contours. Our experiments demonstrate that RNN classifiers trained on complexity contours achieve higher classification accuracy than one trained on text-average complexity scores. In a second step, we determine the relative importance of the features from four distinct categories through a Sensitivity-Based Pruning approach.
Automated writing evaluation systems can improve students’ writing insofar as students attend to the feedback provided and revise their essay drafts in ways aligned with such feedback. Existing research on revision of argumentative writing in such systems, however, has focused on the types of revisions students make (e.g., surface vs. content) rather than the extent to which revisions actually respond to the feedback provided and improve the essay. We introduce an annotation scheme to capture the nature of sentence-level revisions of evidence use and reasoning (the ‘RER’ scheme) and apply it to 5th- and 6th-grade students’ argumentative essays. We show that reliable manual annotation can be achieved and that revision annotations correlate with a holistic assessment of essay improvement in line with the feedback provided. Furthermore, we explore the feasibility of automatically classifying revisions according to our scheme.
Essay traits are attributes of an essay that can help explain how well written (or badly written) the essay is. Examples of traits include Content, Organization, Language, Sentence Fluency, Word Choice, etc. A lot of research in the last decade has dealt with automatic holistic essay scoring - where a machine rates an essay and gives a score for the essay. However, writers need feedback, especially if they want to improve their writing - which is why trait-scoring is important. In this paper, we show how a deep-learning based system can outperform feature-based machine learning systems, as well as a string kernel system in scoring essay traits.
In this paper we present an NLP-based approach for tracking the evolution of written language competence in L2 Spanish learners using a wide range of linguistic features automatically extracted from students’ written productions. Beyond reporting classification results for different scenarios, we explore the connection between the most predictive features and the teaching curriculum, finding that our set of linguistic features often reflect the explicit instructions that students receive during each course.
We consider the problem of automatically suggesting distractors for multiple-choice cloze questions designed for second-language learners. We describe the creation of a dataset including collecting manual annotations for distractor selection. We assess the relationship between the choices of the annotators and features based on distractors and the correct answers, both with and without the surrounding passage context in the cloze questions. Simple features of the distractor and correct answer correlate with the annotations, though we find substantial benefit to additionally using large-scale pretrained models to measure the fit of the distractor in the context. Based on these analyses, we propose and train models to automatically select distractors, and measure the importance of model components quantitatively.
In undergraduate theses, a good methodology section should describe the series of steps that were followed in performing the research. To assist students in this task, we develop machine-learning models and an app that uses them to provide feedback while students write. We construct an annotated corpus that identifies sentences representing methodological steps and labels when a methodology contains a logical sequence of such steps. We train machine-learning models based on language modeling and lexical features that can identify sentences representing methodological steps with 0.939 f-measure, and identify methodology sections containing a logical sequence of steps with an accuracy of 87%. We incorporate these models into a Microsoft Office Add-in, and show that students who improved their methodologies according to the model feedback received better grades on their methodologies.
Multilingual corpora are difficult to compile and a classroom setting adds pedagogy to the mix of factors which make this data so rich and problematic to classify. In this paper, we set out methodological considerations of using automated speech recognition to build a corpus of teacher speech in an Indonesian language classroom. Our preliminary results (64% word error rate) suggest these tools have the potential to speed data collection in this context. We provide practical examples of our data structure, details of our piloted computer-assisted processes, and fine-grained error analysis. Our study is informed and directed by genuine research questions and discussion in both the education and computational linguistics fields. We highlight some of the benefits and risks of using these emerging technologies to analyze the complex work of language teachers and in education more generally.
With the widespread adoption of the Next Generation Science Standards (NGSS), science teachers and online learning environments face the challenge of evaluating students’ integration of different dimensions of science learning. Recent advances in representation learning in natural language processing have proven effective across many natural language processing tasks, but a rigorous evaluation of the relative merits of these methods for scoring complex constructed response formative assessments has not previously been carried out. We present a detailed empirical investigation of feature-based, recurrent neural network, and pre-trained transformer models on scoring content in real-world formative assessment data. We demonstrate that recent neural methods can rival or exceed the performance of feature-based methods. We also provide evidence that different classes of neural models take advantage of different learning cues, and pre-trained transformer models may be more robust to spurious, dataset-specific learning cues, better reflecting scoring rubrics.
We present a computational exploration of argument critique writing by young students. Middle school students were asked to criticize an argument presented in the prompt, focusing on identifying and explaining the reasoning flaws. This task resembles an established college-level argument critique task. Lexical and discourse features that utilize detailed domain knowledge to identify critiques exist for the college task but do not perform well on the young students’ data. Instead, transformer-based architecture (e.g., BERT) fine-tuned on a large corpus of critique essays from the college task performs much better (over 20% improvement in F1 score). Analysis of the performance of various configurations of the system suggests that while children’s writing does not exhibit the standard discourse structure of an argumentative essay, it does share basic local sequential structures with the more mature writers.
Most natural language processing research now recommends large Transformer-based models with fine-tuning for supervised classification tasks; older strategies like bag-of-words features and linear models have fallen out of favor. Here we investigate whether, in automated essay scoring (AES) research, deep neural models are an appropriate technological choice. We find that fine-tuning BERT produces similar performance to classical models at significant additional cost. We argue that while state-of-the-art strategies do match existing best results, they come with opportunity costs in computational resources. We conclude with a review of promising areas for research on student essays where the unique characteristics of Transformers may provide benefits over classical methods to justify the costs.
In this paper, we present a simple and efficient GEC sequence tagger using a Transformer encoder. Our system is pre-trained on synthetic data and then fine-tuned in two stages: first on errorful corpora, and second on a combination of errorful and error-free parallel corpora. We design custom token-level transformations to map input tokens to target corrections. Our best single-model/ensemble GEC tagger achieves an F_0.5 of 65.3/66.5 on CONLL-2014 (test) and F_0.5 of 72.4/73.6 on BEA-2019 (test). Its inference speed is up to 10 times as fast as a Transformer-based seq2seq GEC system.
Complex Word Identification (CWI) is a task for the identification of words that are challenging for second-language learners to read. Even though the use of neural classifiers is now common in CWI, the interpretation of their parameters remains difficult. This paper analyzes neural CWI classifiers and shows that some of their parameters can be interpreted as vocabulary size. We present a novel formalization of vocabulary size measurement methods that are practiced in the applied linguistics field as a kind of neural classifier. We also contribute to building a novel dataset for validating vocabulary testing and readability via crowdsourcing.
Many clinical assessment instruments used to diagnose language impairments in children include a task in which the subject must formulate a sentence to describe an image using a specific target word. Because producing sentences in this way requires the speaker to integrate syntactic and semantic knowledge in a complex manner, responses are typically evaluated on several different dimensions of appropriateness yielding a single composite score for each response. In this paper, we present a dataset consisting of non-clinically elicited responses for three related sentence formulation tasks, and we propose an approach for automatically evaluating their appropriateness. We use neural machine translation to generate correct-incorrect sentence pairs in order to create synthetic data to increase the amount and diversity of training data for our scoring model. Our scoring model uses transfer learning to facilitate automatic sentence appropriateness evaluation. We further compare custom word embeddings with pre-trained contextualized embeddings serving as features for our scoring model. We find that transfer learning improves scoring accuracy, particularly when using pretrained contextualized embeddings.
The tasks of automatically scoring either textual or algebraic responses to mathematical questions have both been well-studied, albeit separately. In this paper we propose a method for automatically scoring responses that contain both text and algebraic expressions. Our method not only achieves high agreement with human raters, but also links explicitly to the scoring rubric – essentially providing explainable models and a way to potentially provide feedback to students in the future.
This paper investigates whether transfer learning can improve the prediction of the difficulty and response time parameters for 18,000 multiple-choice questions from a high-stakes medical exam. The type the signal that best predicts difficulty and response time is also explored, both in terms of representation abstraction and item component used as input (e.g., whole item, answer options only, etc.). The results indicate that, for our sample, transfer learning can improve the prediction of item difficulty when response time is used as an auxiliary task but not the other way around. In addition, difficulty was best predicted using signal from the item stem (the description of the clinical case), while all parts of the item were important for predicting the response time.
Grammatical Error Correction (GEC) is concerned with correcting grammatical errors in written text. Current GEC systems, namely those leveraging statistical and neural machine translation, require large quantities of annotated training data, which can be expensive or impractical to obtain. This research compares techniques for generating synthetic data utilized by the two highest scoring submissions to the restricted and low-resource tracks in the BEA-2019 Shared Task on Grammatical Error Correction.
Gender bias in biomedical research can have an adverse impact on the health of real people. For example, there is evidence that heart disease-related funded research generally focuses on men. Health disparities can form between men and at-risk groups of women (i.e., elderly and low-income) if there is not an equal number of heart disease-related studies for both genders. In this paper, we study temporal bias in biomedical research articles by measuring gender differences in word embeddings. Specifically, we address multiple questions, including, How has gender bias changed over time in biomedical research, and what health-related concepts are the most biased? Overall, we find that traditional gender stereotypes have reduced over time. However, we also find that the embeddings of many medical conditions are as biased today as they were 60 years ago (e.g., concepts related to drug addiction and body dysmorphia).
Novel contexts, comprising a set of terms referring to one or more concepts, may often arise in complex querying scenarios such as in evidence-based medicine (EBM) involving biomedical literature. These may not explicitly refer to entities or canonical concept forms occurring in a fact-based knowledge source, e.g. the UMLS ontology. Moreover, hidden associations between related concepts meaningful in the current context, may not exist within a single document, but across documents in the collection. Predicting semantic concept tags of documents can therefore serve to associate documents related in unseen contexts, or categorize them, in information filtering or retrieval scenarios. Thus, inspired by the success of sequence-to-sequence neural models, we develop a novel sequence-to-set framework with attention, for learning document representations in a unique unsupervised setting, using no human-annotated document labels or external knowledge resources and only corpus-derived term statistics to drive the training, that can effect term transfer within a corpus for semantically tagging a large collection of documents. Our sequence-to-set modeling approach to predict semantic tags, gives to the best of our knowledge, the state-of-the-art for both, an unsupervised query expansion (QE) task for the TREC CDS 2016 challenge dataset when evaluated on an Okapi BM25–based document retrieval system; and also over the MLTM system baseline baseline (Soleimani and Miller, 2016), for both supervised and semi-supervised multi-label prediction tasks on the del.icio.us and Ohsumed datasets. We make our code and data publicly available.
We present a system that allows life-science researchers to search a linguistically annotated corpus of scientific texts using patterns over dependency graphs, as well as using patterns over token sequences and a powerful variant of boolean keyword queries. In contrast to previous attempts to dependency-based search, we introduce a light-weight query language that does not require the user to know the details of the underlying linguistic representations, and instead to query the corpus by providing an example sentence coupled with simple markup. Search is performed at an interactive speed due to efficient linguistic graph-indexing and retrieval engine. This allows for rapid exploration, development and refinement of user queries. We demonstrate the system using example workflows over two corpora: the PubMed corpus including 14,446,243 PubMed abstracts and the CORD-19 dataset, a collection of over 45,000 research papers focused on COVID-19 research. The system is publicly available at https://allenai.github.io/spike
Inferring the nature of the relationships between biomedical entities from text is an important problem due to the difficulty of maintaining human-curated knowledge bases in rapidly evolving fields. Neural word embeddings have earned attention for an apparent ability to encode relational information. However, word embedding models that disregard syntax during training are limited in their ability to encode the structural relationships fundamental to cognitive theories of analogy. In this paper, we demonstrate the utility of encoding dependency structure in word embeddings in a model we call Embedding of Structural Dependencies (ESD) as a way to represent biomedical relationships in two analogical retrieval tasks: a relationship retrieval (RR) task, and a literature-based discovery (LBD) task meant to hypothesize plausible relationships between pairs of entities unseen in training. We compare our model to skip-gram with negative sampling (SGNS), using 19 databases of biomedical relationships as our evaluation data, with improvements in performance on 17 (LBD) and 18 (RR) of these sets. These results suggest embeddings encoding dependency path information are of value for biomedical analogy retrieval.
Improving the quality of medical research reporting is crucial to reduce avoidable waste in research and to improve the quality of health care. Despite various initiatives aiming at improving research reporting – guidelines, checklists, authoring aids, peer review procedures, etc. – overinterpretation of research results, also known as spin, is still a serious issue in research reporting. In this paper, we propose a Natural Language Processing (NLP) system for detecting several types of spin in biomedical articles reporting randomized controlled trials (RCTs). We use a combination of rule-based and machine learning approaches to extract important information on trial design and to detect potential spin. The proposed spin detection system includes algorithms for text structure analysis, sentence classification, entity and relation extraction, semantic similarity assessment. Our algorithms achieved operational performance for the these tasks, F-measure ranging from 79,42 to 97.86% for different tasks. The most difficult task is extracting reported outcomes. Our tool is intended to be used as a semi-automated aid tool for assisting both authors and peer reviewers to detect potential spin. The tool incorporates a simple interface that allows to run the algorithms and visualize their output. It can also be used for manual annotation and correction of the errors in the outputs. The proposed tool is the first tool for spin detection. The tool and the annotated dataset are freely available.
Current research in machine learning for radiology is focused mostly on images. There exists limited work in investigating intelligent interactive systems for radiology. To address this limitation, we introduce a realistic and information-rich task of Visual Dialog in radiology, specific to chest X-ray images. Using MIMIC-CXR, an openly available database of chest X-ray images, we construct both a synthetic and a real-world dataset and provide baseline scores achieved by state-of-the-art models. We show that incorporating medical history of the patient leads to better performance in answering questions as opposed to conventional visual question answering model which looks only at the image. While our experiments show promising results, they indicate that the task is extremely challenging with significant scope for improvement. We make both the datasets (synthetic and gold standard) and the associated code publicly available to the research community.
Recently BERT has achieved a state-of-the-art performance in temporal relation extraction from clinical Electronic Medical Records text. However, the current approach is inefficient as it requires multiple passes through each input sequence. We extend a recently-proposed one-pass model for relation classification to a one-pass model for relation extraction. We augment this framework by introducing global embeddings to help with long-distance relation inference, and by multi-task learning to increase model performance and generalizability. Our proposed model produces results on par with the state-of-the-art in temporal relation extraction on the THYME corpus and is much “greener” in computational cost.
Clinical coding is currently a labour-intensive, error-prone, but a critical administrative process whereby hospital patient episodes are manually assigned codes by qualified staff from large, standardised taxonomic hierarchies of codes. Automating clinical coding has a long history in NLP research and has recently seen novel developments setting new benchmark results. A popular dataset used in this task is MIMIC-III, a large database of clinical free text notes and their associated codes amongst other data. We argue for the reconsideration of the validity MIMIC-III’s assigned codes, as MIMIC-III has not undergone secondary validation. This work presents an open-source, reproducible experimental methodology for assessing the validity of EHR discharge summaries. We exemplify the methodology with MIMIC-III discharge summaries and show the most frequently assigned codes in MIMIC-III are undercoded up to 35%.
Text classification tasks which aim at harvesting and/or organizing information from electronic health records are pivotal to support clinical and translational research. However these present specific challenges compared to other classification tasks, notably due to the particular nature of the medical lexicon and language used in clinical records. Recent advances in embedding methods have shown promising results for several clinical tasks, yet there is no exhaustive comparison of such approaches with other commonly used word representations and classification models. In this work, we analyse the impact of various word representations, text pre-processing and classification algorithms on the performance of four different text classification tasks. The results show that traditional approaches, when tailored to the specific language and structure of the text inherent to the classification task, can achieve or exceed the performance of more recent ones based on contextual embeddings such as BERT.
This paper presents a reinforcement learning approach to extract noise in long clinical documents for the task of readmission prediction after kidney transplant. We face the challenges of developing robust models on a small dataset where each document may consist of over 10K tokens with full of noise including tabular text and task-irrelevant sentences. We first experiment four types of encoders to empirically decide the best document representation, and then apply reinforcement learning to remove noisy text from the long documents, which models the noise extraction process as a sequential decision problem. Our results show that the old bag-of-words encoder outperforms deep learning-based encoders on this task, and reinforcement learning is able to improve upon baseline while pruning out 25% text segments. Our analysis depicts that reinforcement learning is able to identify both typical noisy tokens and task-specific noisy text.
In this paper, we apply pre-trained language models to the Semantic Textual Similarity (STS) task, with a specific focus on the clinical domain. In low-resource setting of clinical STS, these large models tend to be impractical and prone to overfitting. Building on BERT, we study the impact of a number of model design choices, namely different fine-tuning and pooling strategies. We observe that the impact of domain-specific fine-tuning on clinical STS is much less than that in the general domain, likely due to the concept richness of the domain. Based on this, we propose two data augmentation techniques. Experimental results on N2C2-STS 1 demonstrate substantial improvements, validating the utility of the proposed methods.
We explore state-of-the-art neural models for question answering on electronic medical records and improve their ability to generalize better on previously unseen (paraphrased) questions at test time. We enable this by learning to predict logical forms as an auxiliary task along with the main task of answer span detection. The predicted logical forms also serve as a rationale for the answer. Further, we also incorporate medical entity information in these models via the ERNIE architecture. We train our models on the large-scale emrQA dataset and observe that our multi-task entity-enriched models generalize to paraphrased questions ~5% better than the baseline BERT model.
How do we most effectively treat a disease or condition? Ideally, we could consult a database of evidence gleaned from clinical trials to answer such questions. Unfortunately, no such database exists; clinical trial results are instead disseminated primarily via lengthy natural language articles. Perusing all such articles would be prohibitively time-consuming for healthcare practitioners; they instead tend to depend on manually compiled systematic reviews of medical literature to inform care. NLP may speed this process up, and eventually facilitate immediate consult of published evidence. The Evidence Inference dataset was recently released to facilitate research toward this end. This task entails inferring the comparative performance of two treatments, with respect to a given outcome, from a particular article (describing a clinical trial) and identifying supporting evidence. For instance: Does this article report that chemotherapy performed better than surgery for five-year survival rates of operable cancers? In this paper, we collect additional annotations to expand the Evidence Inference dataset by 25%, provide stronger baseline models, systematically inspect the errors that these make, and probe dataset quality. We also release an abstract only (as opposed to full-texts) version of the task for rapid model prototyping. The updated corpus, documentation, and code for new baselines and evaluations are available at http://evidence-inference.ebm-nlp.com/.
Alzheimer’s disease (AD)-related global healthcare cost is estimated to be $1 trillion by 2050. Currently, there is no cure for this disease; however, clinical studies show that early diagnosis and intervention helps to extend the quality of life and inform technologies for personalized mental healthcare. Clinical research indicates that the onset and progression of Alzheimer’s disease lead to dementia and other mental health issues. As a result, the language capabilities of patient start to decline. In this paper, we show that machine learning-based unsupervised clustering of and anomaly detection with linguistic biomarkers are promising approaches for intuitive visualization and personalized early stage detection of Alzheimer’s disease. We demonstrate this approach on 10 year’s (1980 to 1989) of President Ronald Reagan’s speech data set. Key linguistic biomarkers that indicate early-stage AD are identified. Experimental results show that Reagan had early onset of Alzheimer’s sometime between 1983 and 1987. This finding is corroborated by prior work that analyzed his interviews using a statistical technique. The proposed technique also identifies the exact speeches that reflect linguistic biomarkers for early stage AD.
We introduceBIOMRC, a large-scale cloze-style biomedical MRC dataset. Care was taken to reduce noise, compared to the previous BIOREAD dataset of Pappas et al. (2018). Experiments show that simple heuristics do not perform well on the new dataset and that two neural MRC models that had been tested on BIOREAD perform much better on BIOMRC, indicating that the new dataset is indeed less noisy or at least that its task is more feasible. Non-expert human performance is also higher on the new dataset compared to BIOREAD, and biomedical experts perform even better. We also introduce a new BERT-based MRC model, the best version of which substantially outperforms all other methods tested, reaching or surpassing the accuracy of biomedical experts in some experiments. We make the new dataset available in three different sizes, also releasing our code, and providing a leaderboard.
Research on analyzing reading patterns of dyslectic children has mainly been driven by classifying dyslexia types offline. We contend that a framework to remedy reading errors inline is more far-reaching and will help to further advance our understanding of this impairment. In this paper, we propose a simple and intuitive neural model to reinstate migrating words that transpire in letter position dyslexia, a visual analysis deficit to the encoding of character order within a word. Introduced by the anagram matrix representation of an input verse, the novelty of our work lies in the expansion from one to a two dimensional context window for training. This warrants words that only differ in the disposition of letters to remain interpreted semantically similar in the embedding space. Subject to the apparent constraints of the self-attention transformer architecture, our model achieved a unigram BLEU score of 40.6 on our reconstructed dataset of the Shakespeare sonnets.
Identifying the reasons for antibiotic administration in veterinary records is a critical component of understanding antimicrobial usage patterns. This informs antimicrobial stewardship programs designed to fight antimicrobial resistance, a major health crisis affecting both humans and animals in which veterinarians have an important role to play. We propose a document classification approach to determine the reason for administration of a given drug, with particular focus on domain adaptation from one drug to another, and instance selection to minimize annotation effort.
Much of biomedical and healthcare data is encoded in discrete, symbolic form such as text and medical codes. There is a wealth of expert-curated biomedical domain knowledge stored in knowledge bases and ontologies, but the lack of reliable methods for learning knowledge representation has limited their usefulness in machine learning applications. While text-based representation learning has significantly improved in recent years through advances in natural language processing, attempts to learn biomedical concept embeddings so far have been lacking. A recent family of models called knowledge graph embeddings have shown promising results on general domain knowledge graphs, and we explore their capabilities in the biomedical domain. We train several state-of-the-art knowledge graph embedding models on the SNOMED-CT knowledge graph, provide a benchmark with comparison to existing methods and in-depth discussion on best practices, and make a case for the importance of leveraging the multi-relational nature of knowledge graphs for learning biomedical knowledge representation. The embeddings, code, and materials will be made available to the community.
When comparing entities extracted by a medical entity recognition system with gold standard annotations over a test set, two types of mismatches might occur, label mismatch or span mismatch. Here we focus on span mismatch and show that its severity can vary from a serious error to a fully acceptable entity extraction due to the subjectivity of span annotations. For a domain-specific BERT-based NER system, we showed that 25% of the errors have the same labels and overlapping span with gold standard entities. We collected expert judgement which shows more than 90% of these mismatches are accepted or partially accepted by the user. Using the training set of the NER system, we built a fast and lightweight entity classifier to approximate the user experience of such mismatches through accepting or rejecting them. The decisions made by this classifier are used to calculate a learning-based F-score which is shown to be a better approximation of a forgiving user’s experience than the relaxed F-score. We demonstrated the results of applying the proposed evaluation metric for a variety of deep learning medical entity recognition models trained with two datasets.
Fact triples are a common form of structured knowledge used within the biomedical domain. As the amount of unstructured scientific texts continues to grow, manual annotation of these texts for the task of relation extraction becomes increasingly expensive. Distant supervision offers a viable approach to combat this by quickly producing large amounts of labeled, but considerably noisy, data. We aim to reduce such noise by extending an entity-enriched relation classification BERT model to the problem of multiple instance learning, and defining a simple data encoding scheme that significantly reduces noise, reaching state-of-the-art performance for distantly-supervised biomedical relation extraction. Our approach further encodes knowledge about the direction of relation triples, allowing for increased focus on relation learning by reducing noise and alleviating the need for joint learning with knowledge graph completion.
Due to the exponential growth of biomedical literature, event and relation extraction are important tasks in biomedical text mining. Most work only focus on relation extraction, and detect a single entity pair mention on a short span of text, which is not ideal due to long sentences that appear in biomedical contexts. We propose an approach to both relation and event extraction, for simultaneously predicting relationships between all mention pairs in a text. We also perform an empirical study to discuss different network setups for this purpose. The best performing model includes a set of multi-head attentions and convolutions, an adaptation of the transformer architecture, which offers self-attention the ability to strengthen dependencies among related elements, and models the interaction between features extracted by multiple attention heads. Experiment results demonstrate that our approach outperforms the state of the art on a set of benchmark biomedical corpora including BioNLP 2009, 2011, 2013 and BioCreative 2017 shared tasks.
Multi-task learning (MTL) has achieved remarkable success in natural language processing applications. In this work, we study a multi-task learning model with multiple decoders on varieties of biomedical and clinical natural language processing tasks such as text similarity, relation extraction, named entity recognition, and text inference. Our empirical results demonstrate that the MTL fine-tuned models outperform state-of-the-art transformer models (e.g., BERT and its variants) by 2.0% and 1.3% in biomedical and clinical domain adaptation, respectively. Pairwise MTL further demonstrates more details about which tasks can improve or decrease others. This is particularly helpful in the context that researchers are in the hassle of choosing a suitable model for new problems. The code and models are publicly available at https://github.com/ncbi-nlp/bluebert.
We here describe line-a-line, a web-based tool for manual annotation of word-alignments in sentence-aligned parallel corpora. The graphical user interface, which builds on a design template from the Jigsaw system for investigative analysis, displays the words from each sentence pair that is to be annotated as elements in two vertical lists. An alignment between two words is annotated by drag-and-drop, i.e. by dragging an element from the left-hand list and dropping it on an element in the right-hand list. The tool indicates that two words are aligned by lines that connect them and by highlighting associated words when the mouse is hovered over them. Line-a-line uses the efmaral library for producing pre-annotated alignments, on which the user can base the manual annotation. The tool is mainly planned to be used on moderately under-resourced languages, for which resources in the form of parallel corpora are scarce. The automatic word-alignment functionality therefore also incorporates information derived from non-parallel resources, in the form of pre-trained multilingual word embeddings from the MUSE library.
The shared task of the 13th Workshop on Building and Using Comparable Corpora was devoted to the induction of bilingual dictionaries from comparable rather than parallel corpora. In this task, for a number of language pairs involving Chinese, English, French, German, Russian and Spanish, the participants were supposed to determine automatically the target language translations of several thousand source language test words of three frequency ranges. We describe here some background, the task definition, the training and test data sets and the evaluation used for ranking the participating systems. We also summarize the approaches used and present the results of the evaluation. In conclusion, the outcome of the competition are the results of a number of systems which provide surprisingly good solutions to the ambitious problem.
In a bid to reach a larger and more diverse audience, Twitter users often post parallel tweets—tweets that contain the same content but are written in different languages. Parallel tweets can be an important resource for developing machine translation (MT) systems among other natural language processing (NLP) tasks. In this paper, we introduce a generic method for collecting parallel tweets. Using this method, we collect a bilingual corpus of English-Arabic parallel tweets and a list of Twitter accounts who post English-Arabictweets regularly. Since our method is generic, it can also be used for collecting parallel tweets that cover less-resourced languages such as Serbian and Urdu. Additionally, we annotate a subset of Twitter accounts with their countries of origin and topic of interest, which provides insights about the population who post parallel tweets. This latter information can also be useful for author profiling tasks.
In this paper, we show how to use bilingual word embeddings (BWE) to automatically create a corresponding table of meaning tags from two dictionaries in one language and examine the effectiveness of the method. To do this, we had a problem: the meaning tags do not always correspond one-to-one because the granularities of the word senses and the concepts are different from each other. Therefore, we regarded the concept tag that corresponds to a word sense the most as the correct concept tag corresponding the word sense. We used two BWE methods, a linear transformation matrix and VecMap. We evaluated the most frequent sense (MFS) method and the corpus concatenation method for comparison. The accuracies of the proposed methods were higher than the accuracy of the random baseline but lower than those of the MFS and corpus concatenation methods. However, because our method utilized the embedding vectors of the word senses, the relations of the sense tags corresponding to concept tags could be examined by mapping the sense embeddings to the vector space of the concept tags. Also, our methods could be performed when we have only concept or word sense embeddings whereas the MFS method requires a parallel corpus and the corpus concatenation method needs two tagged corpora.
We report an experiment aimed at extracting words expressing a specific semantic relation using intersections of word embeddings. In a multilingual frame-based domain model, specific features of a concept are typically described through a set of non-arbitrary semantic relations. In karstology, our domain of choice which we are exploring though a comparable corpus in English and Croatian, karst phenomena such as landforms are usually described through their FORM, LOCATION, CAUSE, FUNCTION and COMPOSITION. We propose an approach to mine words pertaining to each of these relations by using a small number of seed adjectives, for which we retrieve closest words using word embeddings and then use intersections of these neighbourhoods to refine our search. Such cross-language expansion of semantically-rich vocabulary is a valuable aid in improving the coverage of a multilingual knowledge base, but also in exploring differences between languages in their respective conceptualisations of the domain.
In the context of Machine Translation (MT) from-and-to English, Bahasa Indonesia has been considered a low-resource language, and therefore applying Neural Machine Translation (NMT) which typically requires large training dataset proves to be problematic. In this paper, we show otherwise by collecting large, publicly-available datasets from the Web, which we split into several domains: news, religion, general, and conversation, to train and benchmark some variants of transformer-based NMT models across the domains. We show using BLEU that our models perform well across them , outperform the baseline Statistical Machine Translation (SMT) models, and perform comparably with Google Translate. Our datasets (with the standard split for training, validation, and testing), code, and models are available on https://github.com/gunnxx/indonesian-mt-data
This paper describes and evaluates simple techniques for reducing the research space for parallel sentences in monolingual comparable corpora. Initially, when searching for parallel sentences between two comparable documents, all the possible sentence pairs between the documents have to be considered, which introduces a great degree of imbalance between parallel pairs and non-parallel pairs. This is a problem because even with a high performing algorithm, a lot of noise will be present in the extracted results, thus introducing a need for an extensive and costly manual check phase. We work on a manually annotated subset obtained from a French comparable corpus and show how we can drastically reduce the number of sentence pairs that have to be fed to a classifier so that the results can be manually handled.
The task of Bilingual Dictionary Induction (BDI) consists of generating translations for source language words which is important in the framework of machine translation (MT). The aim of the BUCC 2020 shared task is to perform BDI on various language pairs using comparable corpora. In this paper, we present our approach to the task of English-German and English-Russian language pairs. Our system relies on Bilingual Word Embeddings (BWEs) which are often used for BDI when only a small seed lexicon is available making them particularly effective in a low-resource setting. On the other hand, they perform well on high frequency words only. In order to improve the performance on rare words as well, we combine BWE based word similarity with word surface similarity methods, such as orthography In addition to the often used top-n translation method, we experiment with a margin based approach aiming for dynamic number of translations for each source word. We participate in both the open and closed tracks of the shared task and we show improved results of our method compared to simple vector similarity based approaches. Our system was ranked in the top-3 teams and achieved the best results for English-Russian.
This paper describes the TALN/LS2N system participation at the Building and Using Comparable Corpora (BUCC) shared task. We first introduce three strategies: (i) a word embedding approach based on fastText embeddings; (ii) a concatenation approach using both character Skip-Gram and character CBOW models, and finally (iii) a cognates matching approach based on an exact match string similarity. Then, we present the applied strategy for the shared task which consists in the combination of the embeddings concatenation and the cognates matching approaches. The covered languages are French, English, German, Russian and Spanish. Overall, our system mixing embeddings concatenation and perfect cognates matching obtained the best results while compared to individual strategies, except for English-Russian and Russian-English language pairs for which the concatenation approach was preferred.
Natural Language Processing (NLP), is the field of artificial intelligence that gives the computer the ability to interpret, perceive and extract appropriate information from human languages. Contemporary NLP is predominantly a data driven process. It employs machine learning and statistical algorithms to learn language structures from textual corpus. While application of NLP in English, certain European languages such as Spanish, German, etc. and Chinese, Arabic has been tremendous, it is not so, in many Indian languages. There are obvious advantages in creating aligned bilingual and multilingual corpora. Machine translation, cross-lingual information retrieval, content availability and linguistic comparison are a few of the most sought after applications of such parallel corpora. This paper explains and validates a parallel corpus we created for English-Tamil bilingual pair.
This paper presents a deep learning system for the BUCC 2020 shared task: Bilingual dictionary induction from comparable corpora. We have submitted two runs for this shared Task, German (de) and English (en) language pair for “closed track” and Tamil (ta) and English (en) for the “open track”. Our core approach focuses on quantifying the semantics of the language pairs, so that semantics of two different language pairs can be compared or transfer learned. With the advent of word embeddings, it is possible to quantify this. In this paper, we propose a deep learning approach which makes use of the supplied training data, to generate cross-lingual embedding. This is later used for inducting bilingual dictionary from comparable corpora.
The extraction of anglicisms (lexical borrowings from English) is relevant both for lexicographic purposes and for NLP downstream tasks. We introduce a corpus of European Spanish newspaper headlines annotated with anglicisms and a baseline model for anglicism extraction. In this paper we present: (1) a corpus of 21,570 newspaper headlines written in European Spanish annotated with emergent anglicisms and (2) a conditional random field baseline model with handcrafted features for anglicism extraction. We present the newspaper headlines corpus, describe the annotation tagset and guidelines and introduce a CRF model that can serve as baseline for the task of detecting anglicisms. The presented work is a first step towards the creation of an anglicism extractor for Spanish newswire.
Natural Language Inference (NLI) is the task of inferring the logical relationship, typically entailment or contradiction, between a premise and hypothesis. Code-mixing is the use of more than one language in the same conversation or utterance, and is prevalent in multilingual communities all over the world. In this paper, we present the first dataset for code-mixed NLI, in which both the premises and hypotheses are in code-mixed Hindi-English. We use data from Hindi movies (Bollywood) as premises, and crowd-source hypotheses from Hindi-English bilinguals. We conduct a pilot annotation study and describe the final annotation protocol based on observations from the pilot. Currently, the data collected consists of 400 premises in the form of code-mixed conversation snippets and 2240 code-mixed hypotheses. We conduct an extensive analysis to infer the linguistic phenomena commonly observed in the dataset obtained. We evaluate the dataset using a standard mBERT-based pipeline for NLI and report results.
We investigate when is it beneficial to simultaneously learn representations for several tasks, in low-resource settings. For this, we work with noisy user-generated texts in Algerian, a low-resource non-standardised Arabic variety. That is, to mitigate the problem of the data scarcity, we experiment with jointly learning progressively 4 tasks, namely code-switch detection, named entity recognition, spell normalisation and correction, and identifying users’ sentiments. The selection of these tasks is motivated by the lack of labelled data for automatic morpho-syntactic or semantic sequence-tagging tasks for Algerian, in contrast to the case of much multi-task learning for NLP. Our empirical results show that multi-task learning is beneficial for some tasks in particular settings, and that the effect of each task on another, the order of the tasks, and the size of the training data of the task with more data do matter. Moreover, the data augmentation that we performed with no external resources has been shown to be beneficial for certain tasks.
Code-mixed texts are abundant, especially in social media, and poses a problem for NLP tools, which are typically trained on monolingual corpora. In this paper, we explore and evaluate different types of word embeddings for Indonesian–English code-mixed text. We propose the use of code-mixed embeddings, i.e. embeddings trained on code-mixed text. Because large corpora of code-mixed text are required to train embeddings, we describe a method for synthesizing a code-mixed corpus, grounded in literature and a survey. Using sentiment analysis as a case study, we show that code-mixed embeddings trained on synthesized data are at least as good as cross-lingual embeddings and better than monolingual embeddings.
In a multi-lingual and multi-script society such as India, many users resort to code-mixing while typing on social media. While code-mixing has received a lot of attention in the past few years, it has mostly been studied within a single-script scenario. In this work, we present a case study of Hindi-English bilingual Twitter users while considering the nuances that come with the intermixing of different scripts. We present a concise analysis of how scripts and languages interact in communities and cultures where code-mixing is rampant and offer certain insights into the findings. Our analysis shows that both intra-sentential and inter-sentential script-mixing are present on Twitter and show different behavior in different contexts. Examples suggest that script can be employed as a tool for emphasizing certain phrases within a sentence or disambiguating the meaning of a word. Script choice can also be an indicator of whether a word is borrowed or not. We present our analysis along with examples that bring out the nuances of the different cases.
This paper investigates the use of unsupervised cross-lingual embeddings for solving the problem of code-mixed social media text understanding. We specifically investigate the use of these embeddings for a sentiment analysis task for Hinglish Tweets, viz. English combined with (transliterated) Hindi. In a first step, baseline models, initialized with monolingual embeddings obtained from large collections of tweets in English and code-mixed Hinglish, were trained. In a second step, two systems using cross-lingual embeddings were researched, being (1) a supervised classifier and (2) a transfer learning approach trained on English sentiment data and evaluated on code-mixed data. We demonstrate that incorporating cross-lingual embeddings improves the results (F1-score of 0.635 versus a monolingual baseline of 0.616), without any parallel data required to train the cross-lingual embeddings. In addition, the results show that the cross-lingual embeddings not only improve the results in a fully supervised setting, but they can also be used as a base for distant supervision, by training a sentiment model in one of the source languages and evaluating on the other language projected in the same space. The transfer learning experiments result in an F1-score of 0.556, which is almost on par with the supervised settings and speak to the robustness of the cross-lingual embeddings approach.
We present an analysis of semi-supervised acoustic and language model training for English-isiZulu code-switched (CS) ASR using soap opera speech. Approximately 11 hours of untranscribed multilingual speech was transcribed automatically using four bilingual CS transcription systems operating in English-isiZulu, English-isiXhosa, English-Setswana and English-Sesotho. These transcriptions were incorporated into the acoustic and language model training sets. Results showed that the TDNN-F acoustic models benefit from the additional semi-supervised data and that even better performance could be achieved by including additional CNN layers. Using these CNN-TDNN-F acoustic models, a first iteration of semi-supervised training achieved an absolute mixed-language WER reduction of 3.44%, and a further 2.18% after a second iteration. Although the languages in the untranscribed data were unknown, the best results were obtained when all automatically transcribed data was used for training and not just the utterances classified as English-isiZulu. Despite perplexity improvements, the semi-supervised language model was not able to improve the ASR performance.
In this paper, we explore the methods of obtaining parse trees of code-mixed sentences and analyse the obtained trees. Existing work has shown that linguistic theories can be used to generate code-mixed sentences from a set of parallel sentences. We build upon this work, using one of these theories, the Equivalence-Constraint theory to obtain the parse trees of synthetically generated code-mixed sentences and evaluate them with a neural constituency parser. We highlight the lack of a dataset non-synthetic code-mixed constituency parse trees and how it makes our evaluation difficult. To complete our evaluation, we convert a code-mixed dependency parse tree set into “pseudo constituency trees” and find that a parser trained on synthetically generated trees is able to decently parse these as well.
Code-mixed grapheme-to-phoneme (G2P) conversion is a crucial issue for modern speech recognition and synthesis task, but has been seldom investigated in sentence-level in literature. In this study, we construct a system that performs precise and efficient multi-stage code-mixed G2P conversion, for a less studied agglutinative language, Korean. The proposed system undertakes a sentence-level transliteration that is effective in the accurate processing of Korean text. We formulate the underlying philosophy that supports our approach and demonstrate how it fits with the contemporary document.
Understanding expressed sentiment and emotions are two crucial factors in human multimodal language. This paper describes a Transformer-based joint-encoding (TBJE) for the task of Emotion Recognition and Sentiment Analysis. In addition to use the Transformer architecture, our approach relies on a modular co-attention and a glimpse layer to jointly encode one or more modalities. The proposed solution has also been submitted to the ACL20: Second Grand-Challenge on Multimodal Language to be evaluated on the CMU-MOSEI dataset. The code to replicate the presented experiments is open-source .
Despite the recent advances in opinion mining for written reviews, few works have tackled the problem on other sources of reviews. In light of this issue, we propose a multi-modal approach for mining fine-grained opinions from video reviews that is able to determine the aspects of the item under review that are being discussed and the sentiment orientation towards them. Our approach works at the sentence level without the need for time annotations and uses features derived from the audio, video and language transcriptions of its contents.We evaluate our approach on two datasets and show that leveraging the video and audio modalities consistently provides increased performance over text-only baselines, providing evidence these extra modalities are key in better understanding video reviews.
Sentiment Analysis and Emotion Detection in conversation is key in several real-world applications, with an increase in modalities available aiding a better understanding of the underlying emotions. Multi-modal Emotion Detection and Sentiment Analysis can be particularly useful, as applications will be able to use specific subsets of available modalities, as per the available data. Current systems dealing with Multi-modal functionality fail to leverage and capture - the context of the conversation through all modalities, the dependency between the listener(s) and speaker emotional states, and the relevance and relationship between the available modalities. In this paper, we propose an end to end RNN architecture that attempts to take into account all the mentioned drawbacks. Our proposed model, at the time of writing, out-performs the state of the art on a benchmark dataset on a variety of accuracy and regression metrics.
Our senses individually work in a coordinated fashion to express our emotional intentions. In this work, we experiment with modeling modality-specific sensory signals to attend to our latent multimodal emotional intentions and vice versa expressed via low-rank multimodal fusion and multimodal transformers. The low-rank factorization of multimodal fusion amongst the modalities helps represent approximate multiplicative latent signal interactions. Motivated by the work of~(CITATION) and~(CITATION), we present our transformer-based cross-fusion architecture without any over-parameterization of the model. The low-rank fusion helps represent the latent signal interactions while the modality-specific attention helps focus on relevant parts of the signal. We present two methods for the Multimodal Sentiment and Emotion Recognition results on CMU-MOSEI, CMU-MOSI, and IEMOCAP datasets and show that our models have lesser parameters, train faster and perform comparably to many larger fusion-based architectures.
Allowing humans to communicate through natural language with robots requires connections between words and percepts. The process of creating these connections is called symbol grounding and has been studied for nearly three decades. Although many studies have been conducted, not many considered grounding of synonyms and the employed algorithms either work only offline or in a supervised manner. In this paper, a cross-situational learning based grounding framework is proposed that allows grounding of words and phrases through corresponding percepts without human supervision and online, i.e. it does not require any explicit training phase, but instead updates the obtained mappings for every new encountered situation. The proposed framework is evaluated through an interaction experiment between a human tutor and a robot, and compared to an existing unsupervised grounding framework. The results show that the proposed framework is able to ground words through their corresponding percepts online and in an unsupervised manner, while outperforming the baseline framework.
Behavioral cues play a significant part in human communication and cognitive perception. In most professional domains, employee recruitment policies are framed such that both professional skills and personality traits are adequately assessed. Hiring interviews are structured to evaluate expansively a potential employee’s suitability for the position - their professional qualifications, interpersonal skills, ability to perform in critical and stressful situations, in the presence of time and resource constraints, etc. Candidates, therefore, need to be aware of their positive and negative attributes and be mindful of behavioral cues that might have adverse effects on their success. We propose a multimodal analytical framework that analyzes the candidate in an interview scenario and provides feedback for predefined labels such as engagement, speaking rate, eye contact, etc. We perform a comprehensive analysis that includes the interviewee’s facial expressions, speech, and prosodic information, using the video, audio, and text transcripts obtained from the recorded interview. We use these multimodal data sources to construct a composite representation, which is used for training machine learning classifiers to predict the class labels. Such analysis is then used to provide constructive feedback to the interviewee for their behavioral cues and body language. Experimental validation showed that the proposed methodology achieved promising results.
Building multimodal dialogue understanding capabilities situated in the in-cabin context is crucial to enhance passenger comfort in autonomous vehicle (AV) interaction systems. To this end, understanding passenger intents from spoken interactions and vehicle vision systems is an important building block for developing contextual and visually grounded conversational agents for AV. Towards this goal, we explore AMIE (Automated-vehicle Multimodal In-cabin Experience), the in-cabin agent responsible for handling multimodal passenger-vehicle interactions. In this work, we discuss the benefits of multimodal understanding of in-cabin utterances by incorporating verbal/language input together with the non-verbal/acoustic and visual input from inside and outside the vehicle. Our experimental results outperformed text-only baselines as we achieved improved performances for intent detection with multimodal approach.
An artificial intelligence(AI) system should be capable of processing the sensory inputs to extract both task-specific and general information about its environment. However, most of the existing algorithms extract only task specific information. In this work, an innovative approach to address the problem of processing visual sensory data is presented by utilizing convolutional neural network (CNN). It recognizes and represents the physical and semantic nature of the surrounding in both human readable and machine processable format. This work utilizes the image captioning model to capture the semantics of the input image and a modular design to generate a probability distribution for semantic topics. It gives any autonomous system the ability to process visual information in a human-like way and generates more insights which are hardly possible with a conventional algorithm. Here a model and data collection method are proposed.
Deep Neural Networks have been successfully used for the task of Visual Question Answering for the past few years owing to the availability of relevant large scale datasets. However these datasets are created in artificial settings and rarely reflect the real world scenario. Recent research effectively applies these VQA models for answering visual questions for the blind. Despite achieving high accuracy these models appear to be susceptible to variation in input questions.We analyze popular VQA models through the lens of attribution (input’s influence on predictions) to gain valuable insights. Further, We use these insights to craft adversarial attacks which inflict significant damage to these systems with negligible change in meaning of the input questions. We believe this will enhance development of systems more robust to the possible variations in inputs when deployed to assist the visually impaired.
This paper introduces the citizen science platform, LanguageARC, developed within the NIEUW (Novel Incentives and Workflows) project supported by the National Science Foundation under Grant No. 1730377. LanguageARC is a community-oriented online platform bringing together researchers and “citizen linguists” with the shared goal of contributing to linguistic research and language technology development. Like other Citizen Science platforms and projects, LanguageARC harnesses the power and efforts of volunteers who are motivated by the incentives of contributing to science, learning and discovery, and belonging to a community dedicated to social improvement. Citizen linguists contribute language data and judgments by participating in research tasks such as classifying regional accents from audio clips, recording audio of picture descriptions and answering personality questionnaires to create baseline data for NLP research into autism and neurodegenerative conditions. Researchers can create projects on Language ARC without any coding or HTML required using our Project Builder Toolkit.
Language resources are a major ingredient for the advancement of language technologies. Citizen linguistics can help to create language resources and annotate language resources, not only for the improvement of language technologies, such as machine translation but also for the advancement of linguistic research. The (language) resources covered in this article are a corpus related to the Question of the Month project strand, which was initially aimed at co-creation in citizen linguistics and a partially annotated database of pictures of written text in different languages found in the public sphere. The number of participants in these project strands differed significantly. Especially those activities that were related to data collection (and analysis) had a significantly higher number of contributions per participant. This especially held true for the activities with (prize) incentives. Nevertheless, the activities of the Question of the Month could reach a higher number of participants, even after the co-creation approach was no longer followed. In addition, the Question of the Month brought research gaps and new knowledge to light and challenged existing paradigms and practices. These are especially important for the advancement of scholarly research. Citizen linguistics can help gather and analyze linguistic data, including language resources, in a short period of time. Thus, it may help increase the access to and availability of language resources.
Labelling, or annotation, is the process by which we assign labels to an item with regards to a task. In some Artificial Intelligence problems, such as Computer Vision tasks, the goal is to obtain objective labels. However, in problems such as text and sentiment analysis, subjective labelling is often required. More so when the sentiment analysis deals with actual emotions instead of polarity (positive/negative) . Scientists employ human experts to create these labels, but it is costly and time consuming. Crowdsourcing enables researchers to utilise non-expert knowledge for scientific tasks. From image analysis to semantic annotation, interested researchers can gather a large sample of answers via crowdsourcing platforms in a timely manner. However, non-expert contributions often need to be thoroughly assessed, particularly so when a task is subjective. Researchers have traditionally used ‘Gold Standard’, ‘Thresholding’ and ‘Majority Voting’ as methods to filter non-expert contributions. We argue that these methods are unsuitable for subjective tasks, such as lexicon acquisition and sentiment analysis. We discuss subjectivity in human centered tasks and present a filtering method that defines quality contributors, based on a set of objectively infused terms in a lexicon acquisition task. We evaluate our method against an established lexicon, the diversity of emotions - i.e. subjectivity- and the exclusion of contributions. Our proposed objective evaluation method can be used to assess contributors in subjective tasks that will provide domain agnostic, quality results, with at least 7% improvement over traditional methods.
Crowdsourcing approaches provide a difficult design challenge for developers. There is a trade-off between the efficiency of the task to be done and the reward given to the user for participating, whether it be altruism, social enhancement, entertainment or money. This paper explores how crowdsourcing and citizen science systems collect data and complete tasks, illustrated by a case study from the online language game-with-a-purpose Phrase Detectives. The game was originally developed to be a constrained interface to prevent player collusion, but subsequently benefited from posthoc analysis of over 76k unconstrained inputs from users. Understanding the interface design and task deconstruction are critical for enabling users to participate in such systems and the paper concludes with a discussion of the idea that social networks can be viewed as form of citizen science platform with both constrained and unconstrained inputs making for a highly complex dataset.
Abstract Meaning Representations (AMRs), a syntax-free representation of phrase semantics are useful for capturing the meaning of a phrase and reflecting the relationship between concepts that are referred to. However, annotating AMRs are time consuming and expensive. The existing annotation process requires expertly trained workers who have knowledge of an extensive set of guidelines for parsing phrases. In this paper, we propose a cost-saving two-step process for the creation of a corpus of AMR-phrase pairs for spatial referring expressions. The first step uses non-specialists to perform simple annotations that can be leveraged in the second step to accelerate the annotation performed by the experts. We hypothesize that our process will decrease the cost per annotation and improve consistency across annotators. Few corpora of spatial referring expressions exist and the resulting language resource will be valuable for referring expression comprehension and generation modeling.
We report on a web-based resource for conducting intercomprehension experiments with native speakers of Slavic languages and present our methods for measuring linguistic distances and asymmetries in receptive multilingualism. Through a website which serves as a platform for online testing, a large number of participants with different linguistic backgrounds can be targeted. A statistical language model is used to measure information density and to gauge how language users master various degrees of (un)intelligibilty. The key idea is that intercomprehension should be better when the model adapted for understanding the unknown language exhibits relatively low average distance and surprisal. All obtained intelligibility scores together with distance and asymmetry measures for the different language pairs and processing directions are made available as an integrated online resource in the form of a Slavic intercomprehension matrix (SlavMatrix).
This study uses crowdsourcing through LanguageARC to collect data on levels of accuracy in the identification of speakers’ ethnicities. Ten participants (5 US; 5 South-East England) classified lexically identical speech stimuli from a corpus of 227 speakers aged 18-33yrs from South-East England into the main “ethnic” groups in Britain: White British, Black British and Asian British. Firstly, the data reveals that there is no significant geographic proximity effect on performance between US and British participants. Secondly, results contribute to recent work suggesting that despite the varying heritages of young, ethnic minority speakers in London, they speak an innovative and emerging variety: Multicultural London English (MLE) (e.g. Cheshire et al., 2011). Countering this, participants found perceptual linguistic differences between speakers of all 3 ethnicities (80.7% accuracy). The highest rate of accuracy (96%) was when identifying the ethnicity of Black British speakers from London whose speech seems to form a distinct, perceptual category. Participants also perform substantially better than chance at identifying Black British and Asian British speakers who are not from London (80% and 60% respectively). This suggests that MLE is not a single, homogeneous variety but instead, there are perceptual linguistic differences by ethnicity which transcend the borders of London.
LanguageARC is a portal that offers citizen linguists opportunities to contribute to language related research. It also provides researchers with infrastructure for easily creating data collection and annotation tasks on the portal and potentially connecting with contributors. This document describes LanguageARC’s main features and operation for researchers interested in creating new projects and or using the resulting data.
This paper will detail how IARPA’s MATERIAL Cross-Language Information Retrieval (CLIR) program investigated certain linguistic parameters to guide language choice, data collection and partitioning, and understand evaluation results. Discerning which linguistic parameters correlated with overall performance enabled the evaluation of progress when different languages were measured, and also was an important factor in determining the most effective CLIR pipeline design, customized to handle language-specific properties deemed necessary to address.
The Machine Translation for English Retrieval of Information in Any Language (MATERIAL) research program, sponsored by the Intelligence Advanced Research Projects Activity (IARPA), focuses on rapid development of end-to-end systems capable of retrieving foreign language speech and text documents relevant to different types of English queries that may be further restricted by domain. Those systems also provide evidence of relevance of the retrieved content in the form of English summaries. The program focuses on Less-Resourced Languages and provides its performer teams very limited amounts of annotated training data. This paper describes the corpora that were created for system development and evaluation for the six languages released by the program to date: Tagalog, Swahili, Somali, Lithuanian, Bulgarian and Pashto. The corpora include build packs to train Machine Translation and Automatic Speech Recognition systems; document sets in three text and three speech genres annotated for domain and partitioned for analysis, development and evaluation; and queries of several types together with corresponding binary relevance judgments against the entire set of documents. The paper also describes a detection metric called Actual Query Weighted Value developed by the program to evaluate end-to-end system performance.
At about the midpoint of the IARPA MATERIAL program in October 2019, an evaluation was conducted on systems’ abilities to find Lithuanian documents based on English queries. Subsequently, both the Lithuanian test collection and results from all three teams were made available for detailed analysis. This paper capitalizes on that opportunity to begin to look at what’s working well at this stage of the program, and to identify some promising directions for future work.
We describe an approach to cross lingual information retrieval that does not rely on explicit translation of either document or query terms. Instead, both queries and documents are mapped into a shared embedding space where retrieval is performed. We discuss potential advantages of the approach in handling polysemy and synonymy. We present a method for training the model, and give details of the model implementation. We present experimental results for two cases: Somali-English and Bulgarian-English CLIR.
Multiple neural language models have been developed recently, e.g., BERT and XLNet, and achieved impressive results in various NLP tasks including sentence classification, question answering and document ranking. In this paper, we explore the use of the popular bidirectional language model, BERT, to model and learn the relevance between English queries and foreign-language documents in the task of cross-lingual information retrieval. A deep relevance matching model based on BERT is introduced and trained by finetuning a pretrained multilingual BERT model with weak supervision, using home-made CLIR training data derived from parallel corpora. Experimental results of the retrieval of Lithuanian documents against short English queries show that our model is effective and outperforms the competitive baseline approaches.
We address the problem of linking related documents across languages in a multilingual collection. We evaluate three diverse unsupervised methods to represent and compare documents: (1) multilingual topic model; (2) cross-lingual document embeddings; and (3) Wasserstein distance.We test the performance of these methods in retrieving news articles in Swedish that are known to be related to a given Finnish article.The results show that ensembles of the methods outperform the stand-alone methods, suggesting that they capture complementary characteristics of the documents
In the IARPA MATERIAL program, information retrieval (IR) is treated as a hard detection problem; the system has to output a single global ranking over all queries, and apply a hard threshold on this global list to come up with all the hypothesized relevant documents. This means that how queries are ranked relative to each other can have a dramatic impact on performance. In this paper, we study such a performance measure, the Average Query Weighted Value (AQWV), which is a combination of miss and false alarm rates. AQWV requires that the same detection threshold is applied to all queries. Hence, detection scores of different queries should be comparable, and, to do that, a score normalization technique (commonly used in keyword spotting from speech) should be used. We describe unsupervised methods for score normalization, which are borrowed from the speech field and adapted accordingly for IR, and demonstrate that they greatly improve AQWV on the task of cross-language information retrieval (CLIR), on three low-resource languages used in MATERIAL. We also present a novel supervised score normalization approach which gives additional gains.
In this paper, we describe a cross-lingual information retrieval (CLIR) system that, given a query in English, and a set of audio and text documents in a foreign language, can return a scored list of relevant documents, and present findings in a summary form in English. Foreign audio documents are first transcribed by a state-of-the-art pretrained multilingual speech recognition model that is finetuned to the target language. For text documents, we use multiple multilingual neural machine translation (MT) models to achieve good translation results, especially for low/medium resource languages. The processed documents and queries are then scored using a probabilistic CLIR model that makes use of the probability of translation from GIZA translation tables and scores from a Neural Network Lexical Translation Model (NNLTM). Additionally, advanced score normalization, combination, and thresholding schemes are employed to maximize the Average Query Weighted Value (AQWV) scores. The CLIR output, together with multiple translation renderings, are selected and translated into English snippets via a summarization model. Our turnkey system is language agnostic and can be quickly trained for a new low-resource language in few days.
We describe the human triage scenario envisioned in the Cross-Lingual Information Retrieval (CLIR) problem of the [REDUCT] Program. The overall goal is to maximize the quality of the set of documents that is given to a bilingual analyst, as measured by the AQWV score. The initial set of source documents that are retrieved by the CLIR system is summarized in English and presented to human judges who attempt to remove the irrelevant documents (false alarms); the resulting documents are then presented to the analyst. First, we describe the AQWV performance measure and show that, in our experience, if the acceptance threshold of the CLIR component has been optimized to maximize AQWV, the loss in AQWV due to false alarms is relatively constant across many conditions, which also limits the possible gain that can be achieved by any post filter (such as human judgments) that removes false alarms. Second, we analyze the likely benefits for the triage operation as a function of the initial CLIR AQWV score and the ability of the human judges to remove false alarms without removing relevant documents. Third, we demonstrate that we can increase the benefit for human judgments by combining the human judgment scores with the original document scores returned by the automatic CLIR system.
We describe work from our investigations of the novel area of multi-modal cross-lingual retrieval (MMCLIR) under low-resource conditions. We study the challenges associated with MMCLIR relating to: (i) data conversion between different modalities, for example speech and text, (ii) overcoming the language barrier between source and target languages; (iii) effectively scoring and ranking documents to suit the retrieval task; and (iv) handling low resource constraints that prohibit development of heavily tuned machine translation (MT) and automatic speech recognition (ASR) systems. We focus on the use case of retrieving text and speech documents in Swahili, using English queries which was the main focus of the OpenCLIR shared task. Our work is developed within the scope of this task. In this paper we devote special attention to the automatic translation (AT) component which is crucial for the overall quality of the MMCLIR system. We exploit a combination of dictionaries and phrase-based statistical machine translation (MT) systems to tackle effectively the subtask of query translation. We address each MMCLIR challenge individually, and develop separate components for automatic translation (AT), speech processing (SP) and information retrieval (IR). We find that results with respect to cross-lingual text retrieval are quite good relative to the task of cross-lingual speech retrieval. Overall we find that the task of MMCLIR and specifically cross-lingual speech retrieval is quite complex. Further we pinpoint open issues related to handling cross-lingual audio and text retrieval for low resource languages that need to be addressed in future research.
In this work, we focus on improving ASR output segmentation in the context of low-resource language speech-to-text translation. ASR output segmentation is crucial, as ASR systems segment the input audio using purely acoustic information and are not guaranteed to output sentence-like segments. Since most MT systems expect sentences as input, feeding in longer unsegmented passages can lead to sub-optimal performance. We explore the feasibility of using datasets of subtitles from TV shows and movies to train better ASR segmentation models. We further incorporate part-of-speech (POS) tag and dependency label information (derived from the unsegmented ASR outputs) into our segmentation model. We show that this noisy syntactic information can improve model accuracy. We evaluate our models intrinsically on segmentation quality and extrinsically on downstream MT performance, as well as downstream tasks including cross-lingual information retrieval (CLIR) tasks and human relevance assessments. Our model shows improved performance on downstream tasks for Lithuanian and Bulgarian.
This paper addresses long-term archival for large corpora. Three aspects specific to language resources are focused, namely (1) the removal of resources for legal reasons, (2) versioning of (unchanged) objects in constantly growing resources, especially where objects can be part of multiple releases but also part of different collections, and (3) the conversion of data to new formats for digital preservation. It is motivated why language resources may have to be changed, and why formats may need to be converted. As a solution, the use of an intermediate proxy object called a signpost is suggested. The approach will be exemplified with respect to the corpora of the Leibniz Institute for the German Language in Mannheim, namely the German Reference Corpus (DeReKo) and the Archive for Spoken German (AGD).
We evaluate a graph-based dependency parser on DeReKo, a large corpus of contemporary German. The dependency parser is trained on the German dataset from the SPMRL 2014 Shared Task which contains text from the news domain, whereas DeReKo also covers other domains including fiction, science, and technology. To avoid the need for costly manual annotation of the corpus, we use the parser’s probability estimates for unlabeled and labeled attachment as main evaluation criterion. We show that these probability estimates are highly correlated with the actual attachment scores on a manually annotated test set. On this basis, we compare estimated parsing scores for the individual domains in DeReKo, and show that the scores decrease with increasing distance of a domain to the training corpus.
This paper investigates the impact of different types and size of training corpora on language models. By asking the fundamental question of quality versus quantity, we compare four French corpora by pre-training four different ELMos and evaluating them on dependency parsing, POS-tagging and Named Entities Recognition downstream tasks. We present and asses the relevance of a new balanced French corpus, CaBeRnet, that features a representative range of language usage, including a balanced variety of genres (oral transcriptions, newspapers, popular magazines, technical reports, fiction, academic texts), in oral and written styles. We hypothesize that a linguistically representative corpus will allow the language models to be more efficient, and therefore yield better evaluation scores on different evaluation sets and tasks. This paper offers three main contributions: (1) two newly built corpora: (a) CaBeRnet, a French Balanced Reference Corpus and (b) CBT-fr a domain-specific corpus having both oral and written style in youth literature, (2) five versions of ELMo pre-trained on differently built corpora, and (3) a whole array of computational results on downstream tasks that deepen our understanding of the effects of corpus balance and register in NLP evaluation.
This paper describes work in progress on devising automatic and parallel methods for geoparsing large digital historical textual data by combining the strengths of three natural language processing (NLP) tools, the Edinburgh Geoparser, spaCy and defoe, and employing different tokenisation and named entity recognition (NER) techniques. We apply these tools to a large collection of nineteenth century Scottish geographical dictionaries, and describe preliminary results obtained when processing this data.
Development of dozens of specialized corpus query systems and languages over the past decades has let to a diverse but also fragmented landscape. Today we are faced with a plethora of query tools that each provide unique features, but which are also not interoperable and often rely on very specific database back-ends or formats for storage. This severely hampers usability both for end users that want to query different corpora and also for corpus designers that wish to provide users with an interface for querying and exploration. We propose a hybrid corpus query architecture as a first step to overcoming this issue. It takes the form of a middleware system between user front-ends and optional database or text indexing solutions as back-ends. At its core is a custom query evaluation engine for index-less processing of corpus queries. With a flexible JSON-LD query protocol the approach allows communication with back-end systems to partially solve queries and offset some of the performance penalties imposed by the custom evaluation engine. This paper outlines the details of our first draft of aforementioned architecture.
As a part of the ZuMult-project, we are currently modelling a backend architecture that should provide query access to corpora from the Archive of Spoken German (AGD) at the Leibniz-Institute for the German Language (IDS). We are exploring how to reuse existing search engine frameworks providing full text indices and allowing to query corpora by one of the corpus query languages (QLs) established and actively used in the corpus research community. For this purpose, we tested MTAS - an open source Lucene-based search engine for querying on text with multilevel annotations. We applied MTAS on three oral corpora stored in the TEI-based ISO standard for transcriptions of spoken language (ISO 24624:2016). These corpora differ from the corpus data that MTAS was developed for, because they include interactions with two and more speakers and are enriched, inter alia, with timeline-based annotations. In this contribution, we report our test results and address issues that arise when search frameworks originally developed for querying written corpora are being transferred into the field of spoken language.
The challenges for making use of a large text corpus such as the ‘AAC – Austrian Academy Corpus’ for the purposes of digital literary studies will be addressed in this presentation. The research question of how to use a digital text corpus of considerable size for such a specific research purpose is of interest for corpus research in general as it is of interest for digital literary text studies which rely to a large extent on large digital text corpora. The observations of the usage of lexical entities such as words, word forms, multi word units and many other linguistic units determine the way in which texts are being studied and explored. Larger entities have to be taken into account as well, which is why questions of semantic analysis and larger structures come into play. The texts of the AAC – Austrian Academy Corpus which was founded in 2001 are German language texts of historical and cultural significance from the time between 1848 and 1989. The aim of this study is to present possible research questions for corpus-based methodological approaches for the digital study of literary texts and to give examples of early experiments and experiences with making use of a large text corpus for these research purposes.
The paper overviews the state of implementation of the Czech National Corpus (CNC) in all the main areas of its operation: corpus compilation, annotation, application development and user services. As the focus is on the recent development, some of the areas are described in more detail than the others. Close attention is paid to the data collection and, in particular, to the description of web application development. This is not only because CNC has recently seen a significant progress in this area, but also because we believe that end-user web applications shape the way linguists and other scholars think about the language data and about the range of possibilities they offer. This consideration is even more important given the variability of the CNC corpora.
In this paper we present an experiment of augmenting the Corpus of Contemporary Romanian Language (CoRoLa) with the syntactic level of annotations, which would allow users to address queries about the syntax of Romanian sentences, in the Universal Dependency model. After a short introduction of CoRoLa, we describe the treebanks used to train the dependency parser, we show the evaluation results and the process of upgrading CoRoLa with the new level of annotations. The parser displaying the best accuracy with respect to recognition of heads and relations, out of three variants trained on manually built treebanks, was chosen. Keywords: Syntactic annotation, treebank, corpus, maltparser
The first step of any terminological work is to setup a reliable, specialized corpus composed of documents written by specialists and then to apply automatic term extraction (ATE) methods to this corpus in order to retrieve a first list of potential terms. In this paper, the experiment we describe differs quite drastically from this usual process since we are applying ATE to unspecialized corpora. The corpus used for this study was built from newspaper articles retrieved from the Web using a short list of keywords. The general intuition on which this research is based is that ATE based corpus comparison techniques can be used to capture both similarities and dissimilarities between corpora. The former are exploited through a termhood measure and the latter through word embeddings. Our initial results were validated manually and show that combining a traditional ATE method that focuses on dissimilarities between corpora to newer methods that exploit similarities (more specifically distributional features of candidates) leads to promising results.
Automatic term extraction (ATE) from texts is critical for effective terminology work in small speech communities. We present TermPortal, a workbench for terminology work in Iceland, featuring the first ATE system for Icelandic. The tool facilitates standardization in terminology work in Iceland, as it exports data in standard formats in order to streamline gathering and distribution of the material. In the project we focus on the domain of finance in order to do be able to fulfill the needs of an important and large field. We present a comprehensive survey amongst the most prominent organizations in that field, the results of which emphasize the need for a good, up-to-date and accessible termbank and the willingness to use terms in Icelandic. Furthermore we present the ATE tool for Icelandic, which uses a variety of methods and shows great potential with a recall rate of up to 95% and a high C-value, indicating that it competently finds term candidates that are important to the input text.
A common method of structuring information extracted from textual data is using a knowledge model (e.g. a thesaurus) to organise the information semantically. Creating and managing a knowledge model is already a costly task in terms of human effort, not to mention making it multilingual. Multilingual knowledge modelling is a common problem for both transnational organisations and organisations providing text analytics that want to analyse information in more than one language. Many organisations tend to develop their language resources first in one language (often English). When it comes to analysing data sources in other languages, either a lot of effort has to be invested in recreating the same knowledge base in a different language or the data itself has to be translated into the language of the knowledge model. In this paper, we propose an unsupervised method to automatically induce a given thesaurus into another language using only comparable monolingual corpora. The aim of this proposal is to employ cross-lingual word embeddings to map the set of topics in an already-existing English thesaurus into Spanish. With this in mind, we describe different approaches to generate the Spanish thesaurus terms and offer an extrinsic evaluation by using the obtained thesaurus, which covers non-financial topics in a multi-label document classification task, and we compare the results across these approaches.
We present a study whose objective is to compare several dependency parsers for English applied to a specialized corpus for building distributional count-based models from syntactic dependencies. One of the particularities of this study is to focus on the concepts of the target domain, which mainly occur in documents as multi-terms and must be aligned with the outputs of the parsers. We compare a set of ten parsers in terms of syntactic triplets but also in terms of distributional neighbors extracted from the models built from these triplets, both with and without an external reference concerning the semantic relations between concepts. We show more particularly that some patterns of proximity between these parsers can be observed across our different evaluations, which could give insights for anticipating the performance of a parser for building distributional models from a given corpus
Machine learning plays an ever-bigger part in online recruitment, powering intelligent matchmaking and job recommendations across many of the world’s largest job platforms. However, the main text is rarely enough to fully understand a job posting: more often than not, much of the required information is condensed into the job title. Several organised efforts have been made to map job titles onto a hand-made knowledge base as to provide this information, but these only cover around 60% of online vacancies. We introduce a novel, purely data-driven approach towards the detection of new job titles. Our method is conceptually simple, extremely efficient and competitive with traditional NER-based approaches. Although the standalone application of our method does not outperform a finetuned BERT model, it can be applied as a preprocessing step as well, substantially boosting accuracy across several architectures.
The empowerment of the population and the democratisation of information regarding healthcare have revealed that there is a communication gap between health professionals and patients. The latter are constantly receiving more and more written information about their healthcare visits and treatments, but that does not mean they understand it. In this paper we focus on the patient’s lack of comprehension of medical reports. After linguistically characterising the medical report, we present the results of a survey that showed that patients have serious comprehension difficulties concerning the medical reports they receive, specifically problems regarding the medical terminology used in these texts, specifically in Spanish and Catalan. To favour the understanding of medical reports, we propose an automatic text enrichment strategy that generates linguistically and cognitively enriched medical reports which are more comprehensible to the patient, and which focus on the parts of the medical report that most interest the patient: the diagnosis and treatment sections.
The semantic projection method is often used in terminology structuring to infer semantic relations between terms. Semantic projection relies upon the assumption of semantic compositionality: the relation that links simple term pairs remains valid in pairs of complex terms built from these simple terms. This paper proposes to investigate whether this assumption commonly adopted in natural language processing is actually valid. First, we describe the process of constructing a list of semantically linked multi-word terms (MWTs) related to the environmental field through the extraction of semantic variants. Second, we present our analysis of the results from the semantic projection. We find that contexts play an essential role in defining the relations between MWTs.
We present the NetViz terminology visualization tool and apply it to the domain modeling of karstology, a subfield of geography studying karst phenomena. The developed tool allows for high-performance online network visualization where the user can upload the terminological data in a simple CSV format, define the nodes (terms, categories), edges (relations) and their properties (by assigning different node colors), and then edit and interactively explore domain knowledge in the form of a network. We showcase the usefulness of the tool on examples from the karstology domain, where in the first use case we visualize the domain knowledge as represented in a manually annotated corpus of domain definitions, while in the second use case we show the power of visualization for domain understanding by visualizing automatically extracted knowledge in the form of triplets extracted from the karstology domain corpus. The application is entirely web-based without any need for downloading or special configuration. The source code of the web application is also available under the permissive MIT license, allowing future extensions for developing new terminological applications.
Thesaurus construction with minimum human efforts often relies on automatic methods to discover terms and their relations. Hence, the quality of a thesaurus heavily depends on the chosen methodologies for: (i) building its content (terminology extraction task) and (ii) designing its structure (semantic similarity task). The performance of the existing methods on automatic thesaurus construction is still less accurate than the handcrafted ones of which is important to highlight the drawbacks to let new strategies build more accurate thesauri models. In this paper, we will provide a systematic analysis of existing methods for both tasks and discuss their feasibility based on an Italian Cybersecurity corpus. In particular, we will provide a detailed analysis on how the semantic relationships network of a thesaurus can be automatically built, and investigate the ways to enrich the terminological scope of a thesaurus by taking into account the information contained in external domain-oriented semantic sets.
Terminology extraction procedure usually consists of selecting candidates for terms and ordering them according to their importance for the given text or set of texts. Depending on the method used, a list of candidates contains different fractions of grammatically incorrect, semantically odd and irrelevant sequences. The aim of this work was to improve term candidate selection by reducing the number of incorrect sequences using a dependency parser for Polish.
Our contribution is part of a wider research project on term variation in German and concentrates on the computational aspects of a frame-based model for term meaning representation in the technical field. We focus on the role of frames (in the sense of Frame-Based Terminology) as the semantic interface between concepts covered by a domain ontology and domain-specific terminology. In particular, we describe methods for performing frame-based corpus annotation and frame-based term extraction. The aim of the contribution is to discuss the capacity of the model to automatically acquire semantic knowledge suitable for terminographic information tools such as specialised dictionaries, and its applicability to further specialised languages.
The TermEval 2020 shared task provided a platform for researchers to work on automatic term extraction (ATE) with the same dataset: the Annotated Corpora for Term Extraction Research (ACTER). The dataset covers three languages (English, French, and Dutch) and four domains, of which the domain of heart failure was kept as a held-out test set on which final f1-scores were calculated. The aim was to provide a large, transparent, qualitatively annotated, and diverse dataset to the ATE research community, with the goal of promoting comparative research and thus identifying strengths and weaknesses of various state-of-the-art methodologies. The results show a lot of variation between different systems and illustrate how some methodologies reach higher precision or recall, how different systems extract different types of terms, how some are exceptionally good at finding rare terms, or are less impacted by term length. The current contribution offers an overview of the shared task with a comparative evaluation, which complements the individual papers by all participants.
Automatic terminology extraction is a notoriously difficult task aiming to ease effort demanded to manually identify terms in domain-specific corpora by automatically providing a ranked list of candidate terms. The main ways that addressed this task can be ranged in four main categories: (i) rule-based approaches, (ii) feature-based approaches, (iii) context-based approaches, and (iv) hybrid approaches. For this first TermEval shared task, we explore a feature-based approach, and a deep neural network multitask approach -BERT- that we fine-tune for term extraction. We show that BERT models (RoBERTa for English and CamemBERT for French) outperform other systems for French and English languages.
This paper describes RACAI’s automatic term extraction system, which participated in the TermEval 2020 shared task on English monolingual term extraction. We discuss the system architecture, some of the challenges that we faced as well as present our results in the English competition.
The identification of terms from domain-specific corpora using computational methods is a highly time-consuming task because terms has to be validated by specialists. In order to improve term candidate selection, we have developed the Token Slot Recognition (TSR) method, a filtering strategy based on terminological tokens which is used to rank extracted term candidates from domain-specific corpora. We have implemented this filtering strategy in TBXTools. In this paper we present the system we have used in the TermEval 2020 shared task on monolingual term extraction. We also present the evaluation results for the system for English, French and Dutch and for two corpora: corruption and heart failure. For English and French we have used a linguistic methodology based on POS patterns, and for Dutch we have used a statistical methodology based on n-grams calculation and filtering with stop-words. For all languages, TSR (Token Slot Recognition) filtering method has been applied. We have obtained competitive results, but there is still room for improvement of the system.
In this work, we introduce a bootstrapped, iterative NER model that integrates a PU learning algorithm for recognizing named entities in a low-resource setting. Our approach combines dictionary-based labeling with syntactically-informed label expansion to efficiently enrich the seed dictionaries. Experimental results on a dataset of manually annotated e-commerce product descriptions demonstrate the effectiveness of the proposed framework.
In an attempt to balance precision and recall in the search page, leading digital shops have been effectively nudging users into select category facets as early as in the type-ahead suggestions. In this work, we present SessionPath, a novel neural network model that improves facet suggestions on two counts: first, the model is able to leverage session embeddings to provide scalable personalization; second, SessionPath predicts facets by explicitly producing a probability distribution at each node in the taxonomy path. We benchmark SessionPath on two partnering shops against count-based and neural models, and show how business requirements and model behavior can be combined in a principled way.
Alternative recommender systems are critical for ecommerce companies. They guide customers to explore a massive product catalog and assist customers to find the right products among an overwhelming number of options. However, it is a non-trivial task to recommend alternative products that fit customers’ needs. In this paper, we use both textual product information (e.g. product titles and descriptions) and customer behavior data to recommend alternative products. Our results show that the coverage of alternative products is significantly improved in offline evaluations as well as recall and precision. The final A/B test shows that our algorithm increases the conversion rate by 12% in a statistically significant way. In order to better capture the semantic meaning of product information, we build a Siamese Network with Bidirectional LSTM to learn product embeddings. In order to learn a similarity space that better matches the preference of real customers, we use co-compared data from historical customer behavior as labels to train the network. In addition, we use NMSLIB to accelerate the computationally expensive kNN computation for millions of products so that the alternative recommendation is able to scale across the entire catalog of a major ecommerce site.
Sentiment analysis is crucial for the advancement of artificial intelligence (AI). Sentiment understanding can help AI to replicate human language and discourse. Studying the formation and response of sentiment state from well-trained Customer Service Representatives (CSRs) can help make the interaction between humans and AI more intelligent. In this paper, a sentiment analysis pipeline is first carried out with respect to real-world multi-party conversations - that is, service calls. Based on the acoustic and linguistic features extracted from the source information, a novel aggregated method for voice sentiment recognition framework is built. Each party’s sentiment pattern during the communication is investigated along with the interaction sentiment pattern between all parties.
While buying a product from the e-commerce websites, customers generally have a plethora of questions. From the perspective of both the e-commerce service provider as well as the customers, there must be an effective question answering system to provide immediate answer to the user queries. While certain questions can only be answered after using the product, there are many questions which can be answered from the product specification itself. Our work takes a first step in this direction by finding out the relevant product specifications, that can help answering the user questions. We propose an approach to automatically create a training dataset for this problem. We utilize recently proposed XLNet and BERT architectures for this problem and find that they provide much better performance than the Siamese model, previously applied for this problem. Our model gives a good performance even when trained on one vertical and tested across different verticals.
In this work, we improve the intent classification in an English based e-commerce voice assistant by using inter-utterance context. For increased user adaptation and hence being more profitable, an e-commerce voice assistant is desired to understand the context of a conversation and not have the users repeat it in every utterance. For example, let a user’s first utterance be ‘find apples’. Then, the user may say ‘i want organic only’ to filter out the results generated by an assistant with respect to the first query. So, it is important for the assistant to take into account the context from the user’s first utterance to understand her intention in the second one. In this paper, we present our approach for contextual intent classification in Walmart’s e-commerce voice assistant. It uses the intent of the previous user utterance to predict the intent of her current utterance. With the help of experiments performed on real user queries we show that our approach improves the intent classification in the assistant.
In this paper, we present a semi-supervised bootstrapping approach to detect product or service related complaints in social media. Our approach begins with a small collection of annotated samples which are used to identify a preliminary set of linguistic indicators pertinent to complaints. These indicators are then used to expand the dataset. The expanded dataset is again used to extract more indicators. This process is applied for several iterations until we can no longer find any new indicators. We evaluated this approach on a Twitter corpus specifically to detect complaints about transportation services. We started with an annotated set of 326 samples of transportation complaints, and after four iterations of the approach, we collected 2,840 indicators and over 3,700 tweets. We annotated a random sample of 700 tweets from the final dataset and observed that nearly half the samples were actual transportation complaints. Lastly, we also studied how different features based on semantics, orthographic properties, and sentiment contribute towards the prediction of complaints.
In e-commerce, recommender systems have become an indispensable part of helping users explore the available inventory. In this work, we present a novel approach for item-based collaborative filtering, by leveraging BERT to understand items, and score relevancy between different items. Our proposed method could address problems that plague traditional recommender systems such as cold start, and “more of the same” recommended content. We conducted experiments on a large-scale real-world dataset with full cold-start scenario, and the proposed approach significantly outperforms the popular Bi-LSTM model.
Product reviews are a huge source of natural language data in e-commerce applications. Several millions of customers write reviews regarding a variety of topics. We categorize these topics into two groups as either “category-specific” topics or as “generic” topics that span multiple product categories. While we can use a supervised learning approach to tag review text for generic topics, it is impossible to use supervised approaches to tag category-specific topics due to the sheer number of possible topics for each category. In this paper, we present an approach to tag each review with several product category-specific tags on Indonesian language product reviews using a semi-supervised approach. We show that our proposed method can work at scale on real product reviews at Tokopedia, a major e-commerce platform in Indonesia. Manual evaluation shows that the proposed method can efficiently generate category-specific product tags.
In e-commerce system, category prediction is to automatically predict categories of given texts. Different from traditional classification where there are no relations between classes, category prediction is reckoned as a standard hierarchical classification problem since categories are usually organized as a hierarchical tree. In this paper, we address hierarchical category prediction. We propose a Deep Hierarchical Classification framework, which incorporates the multi-scale hierarchical information in neural networks and introduces a representation sharing strategy according to the category tree. We also define a novel combined loss function to punish hierarchical prediction losses. The evaluation shows that the proposed approach outperforms existing approaches in accuracy.
In recent years, there has been an increase in online shopping resulting in an increased number of online reviews. Customers cannot delve into the huge amount of data when they are looking for specific aspects of a product. Some of these aspects can be extracted from the product reviews. In this paper we introduced SimsterQ - a clustering based system for answering questions that makes use of word vectors. Clustering was performed using cosine similarity scores between sentence vectors of reviews and questions. Two variants (Sim and Median) with and without stopwords were evaluated against traditional methods that use term frequency. We also used an n-gram approach to study the effect of noise. We used the reviews in the Amazon Reviews dataset to pick the answers. Evaluation was performed both at the individual sentence level using the top sentence from Okapi BM25 as the gold standard and at the whole answer level using review snippets as the gold standard. At the sentence level our system performed slightly better than a more complicated deep learning method. Our system returned answers similar to the review snippets from the Amazon QA Dataset as measured by the cosine similarity. Analysis was also performed on the quality of the clusters generated by our system.
In recent years, the focus of e-Commerce research has been on better understanding the relationship between the internet marketplace, customers, and goods and services. This has been done by examining information that can be gleaned from consumer information, recommender systems, click rates, or the way purchasers go about making buying decisions, for example. This paper takes a very different approach and examines the companies themselves. In the past ten years, e-Commerce giants such as Amazon, Skymall, Wayfair, and Groupon have been embroiled in class action security lawsuits promulgated under Rule 10b(5), which, in short, is one of the Securities and Exchange Commission’s main rules surrounding fraud. Lawsuits are extremely expensive to the company and can damage a company’s brand extensively, with the shareholders left to suffer the consequences. We examined the Management Discussion and Analysis and the Market Risks for 96 companies using sentiment analysis on selected financial measures and found that we were able to predict the outcome of the lawsuits in our dataset using sentiment (tone) alone to a recall of 0.8207 using the Random Forest classifier. We believe that this is an important contribution as it has cross-domain implications and potential, and opens up new areas of research in e-Commerce, finance, and law, as the settlements from the class action lawsuits in our dataset alone are in excess of $1.6 billion dollars, in aggregate.
In this paper, we study the applicability of Bayesian Parametric and Non-parametric methods for user clustering in an E-commerce search setting. To the best of our knowledge, this is the first work that presents a comparative study of various Bayesian clustering methods in the context of product search. Specifically, we cluster users based on their topical patterns from their respective product search queries. To evaluate the quality of the clusters formed, we perform a collaborative query recommendation task. Our findings indicate that simple parametric model like Latent Dirichlet Allocation (LDA) outperforms more sophisticated non-parametric methods like Distance Dependent Chinese Restaurant Process and Dirichlet Process-based clustering in both tasks.
Automatic fact checking is an important task motivated by the need for detecting and preventing the spread of misinformation across the web. The recently released FEVER challenge provides a benchmark task that assesses systems’ capability for both the retrieval of required evidence and the identification of authentic claims. Previous approaches share a similar pipeline training paradigm that decomposes the task into three subtasks, with each component built and trained separately. Although achieving acceptable scores, these methods induce difficulty for practical application development due to unnecessary complexity and expensive computation. In this paper, we explore the potential of simplifying the system design and reducing training computation by proposing a joint training setup in which a single sequence matching model is trained with compounded labels that give supervision for both sentence selection and claim verification subtasks, eliminating the duplicate computation that occurs when models are designed and trained separately. Empirical results on FEVER indicate that our method: (1) outperforms the typical multi-task learning approach, and (2) gets comparable results to top performing systems with a much simpler training setup and less training computation (in terms of the amount of data consumed and the number of model parameters), facilitating future works on the automatic fact checking task and its practical usage.
This work explores the application of textual entailment in news claim verification and stance prediction using a new corpus in Arabic. The publicly available corpus comes in two perspectives: a version consisting of 4,547 true and false claims and a version consisting of 3,786 pairs (claim, evidence). We describe the methodology for creating the corpus and the annotation process. Using the introduced corpus, we also develop two machine learning baselines for two proposed tasks: claim verification and stance prediction. Our best model utilizes pretraining (BERT) and achieves 76.7 F1 on the stance prediction task and 64.3 F1 on the claim verification task. Our preliminary experiments shed some light on the limits of automatic claim verification that relies on claims text only. Results hint that while the linguistic features and world knowledge learned during pretraining are useful for stance prediction, such learned representations from pretraining are insufficient for verifying claims without access to context or evidence.
Textual patterns (e.g., Country’s president Person) are specified and/or generated for extracting factual information from unstructured data. Pattern-based information extraction methods have been recognized for their efficiency and transferability. However, not every pattern is reliable: A major challenge is to derive the most complete and accurate facts from diverse and sometimes conflicting extractions. In this work, we propose a probabilistic graphical model which formulates fact extraction in a generative process. It automatically infers true facts and pattern reliability without any supervision. It has two novel designs specially for temporal facts: (1) it models pattern reliability on two types of time signals, including temporal tag in text and text generation time; (2) it models commonsense constraints as observable variables. Experimental results demonstrate that our model significantly outperforms existing methods on extracting true temporal facts from news data.
In the field of factoid question answering (QA), it is known that the state-of-the-art technology has achieved an accuracy comparable to that of humans in a certain benchmark challenge. On the other hand, in the area of non-factoid QA, there is still a limited number of datasets for training QA models, i.e., machine comprehension models. Considering such a situation within the field of the non-factoid QA, this paper aims to develop a dataset for training Japanese how-to tip QA models. This paper applies one of the state-of-the-art machine comprehension models to the Japanese how-to tip QA dataset. The trained how-to tip QA model is also compared with a factoid QA model trained with a Japanese factoid QA dataset. Evaluation results revealed that the how-to tip machine comprehension performance was almost comparative with that of the factoid machine comprehension even with the training data size reduced to around 4% of the factoid machine comprehension. Thus, the how-to tip machine comprehension task requires much less training data compared with the factoid machine comprehension task.
Recent work has suggested that language models (LMs) store both common-sense and factual knowledge learned from pre-training data. In this paper, we leverage this implicit knowledge to create an effective end-to-end fact checker using a solely a language model, without any external knowledge or explicit retrieval components. While previous work on extracting knowledge from LMs have focused on the task of open-domain question answering, to the best of our knowledge, this is the first work to examine the use of language models as fact checkers. In a closed-book setting, we show that our zero-shot LM approach outperforms a random baseline on the standard FEVER task, and that our finetuned LM compares favorably with standard baselines. Though we do not ultimately outperform methods which use explicit knowledge bases, we believe our exploration shows that this method is viable and has much room for exploration.
We propose two measures for measuring the quality of constructed claims in the FEVER task. Annotating data for this task involves the creation of supporting and refuting claims over a set of evidence. Automatic annotation processes often leave superficial patterns in data, which learning systems can detect instead of performing the underlying task. Humans also can leave these superficial patterns, either voluntarily or involuntarily (due to e.g. fatigue). The two measures introduced attempt to detect the impact of these superficial patterns. One is a new information-theoretic and distributionality based measure, DCI; and the other an extension of neural probing work over the ARCT task, utility. We demonstrate these measures over a recent major dataset, that from the English FEVER task in 2019.
The alarming spread of fake news in social media, together with the impossibility of scaling manual fact verification, motivated the development of natural language processing techniques to automatically verify the veracity of claims. Most approaches perform a claim-evidence classification without providing any insights about why the claim is trustworthy or not. We propose, instead, a model-agnostic framework that consists of two modules: (1) a span extractor, which identifies the crucial information connecting claim and evidence; and (2) a classifier that combines claim, evidence, and the extracted spans to predict the veracity of the claim. We show that the spans are informative for the classifier, improving performance and robustness. Tested on several state-of-the-art models over the Fever dataset, the enhanced classifiers consistently achieve higher accuracy while also showing reduced sensitivity to artifacts in the claims.
Detecting sarcasm and verbal irony is critical for understanding people’s actual sentiments and beliefs. Thus, the field of sarcasm analysis has become a popular research problem in natural language processing. As the community working on computational approaches for sarcasm detection is growing, it is imperative to conduct benchmarking studies to analyze the current state-of-the-art, facilitating progress in this area. We report on the shared task on sarcasm detection we conducted as a part of the 2nd Workshop on Figurative Language Processing (FigLang 2020) at ACL 2020.
We present a novel data augmentation technique, CRA (Contextual Response Augmentation), which utilizes conversational context to generate meaningful samples for training. We also mitigate the issues regarding unbalanced context lengths by changing the input output format of the model such that it can deal with varying context lengths effectively. Specifically, our proposed model, trained with the proposed data augmentation technique, participated in the sarcasm detection task of FigLang2020, have won and achieves the best performance in both Reddit and Twitter datasets.
In this paper, we report on the shared task on metaphor identification on VU Amsterdam Metaphor Corpus and on a subset of the TOEFL Native Language Identification Corpus. The shared task was conducted as apart of the ACL 2020 Workshop on Processing Figurative Language.
Machine metaphor understanding is one of the major topics in NLP. Most of the recent attempts consider it as classification or sequence tagging task. However, few types of research introduce the rich linguistic information into the field of computational metaphor by leveraging powerful pre-training language models. We focus a novel reading comprehension paradigm for solving the token-level metaphor detection task which provides an innovative type of solution for this task. We propose an end-to-end deep metaphor detection model named DeepMet based on this paradigm. The proposed approach encodes the global text context (whole sentence), local text context (sentence fragments), and question (query word) information as well as incorporating two types of part-of-speech (POS) features by making use of the advanced pre-training language model. The experimental results by using several metaphor datasets show that our model achieves competitive results in the second shared task on metaphor detection.
While mysterious, humor likely hinges on an interplay of entities, their relationships, and cultural connotations. Motivated by the importance of context in humor, we consider methods for constructing and leveraging contextual representations in generating humorous text. Specifically, we study the capacity of transformer-based architectures to generate funny satirical headlines, and show that both language models and summarization models can be fine-tuned to regularly generate headlines that people find funny. Furthermore, we find that summarization models uniquely support satire-generation by enabling the generation of topical humorous text. Outside of our formal study, we note that headlines generated by our model were accepted via a competitive process into a satirical newspaper, and one headline was ranked as high or better than 73% of human submissions. As part of our work, we contribute a dataset of over 15K satirical headlines paired with ranked contextual information from news articles and Wikipedia.
Sarcasm is an intricate form of speech, where meaning is conveyed implicitly. Being a convoluted form of expression, detecting sarcasm is an assiduous problem. The difficulty in recognition of sarcasm has many pitfalls, including misunderstandings in everyday communications, which leads us to an increasing focus on automated sarcasm detection. In the second edition of the Figurative Language Processing (FigLang 2020) workshop, the shared task of sarcasm detection released two datasets, containing responses along with their context sampled from Twitter and Reddit. In this work, we use RoBERTalarge to detect sarcasm in both the datasets. We further assert the importance of context in improving the performance of contextual word embedding based models by using three different types of inputs - Response-only, Context-Response, and Context-Response (Separated). We show that our proposed architecture performs competitively for both the datasets. We also show that the addition of a separation token between context and target response results in an improvement of 5.13% in the F1-score in the Reddit dataset.
Sarcasm is a form of communication in which the person states opposite of what he actually means. In this paper, we propose using machine learning techniques with BERT and GloVe embeddings to detect sarcasm in tweets. The dataset is preprocessed before extracting the embeddings. The proposed model also uses all of the context provided in the dataset to which the user is reacting along with his actual response.
Automatic Sarcasm Detection in conversations is a difficult and tricky task. Classifying an utterance as sarcastic or not in isolation can be futile since most of the time the sarcastic nature of a sentence heavily relies on its context. This paper presents our proposed model, C-Net, which takes contextual information of a sentence in a sequential manner to classify it as sarcastic or non-sarcastic. Our model showcases competitive performance in the Sarcasm Detection shared task organised on CodaLab and achieved 75.0% F1-score on the Twitter dataset and 66.3% F1-score on Reddit dataset.
Sarcasm is a type of figurative language broadly adopted in social media and daily conversations. The sarcasm can ultimately alter the meaning of the sentence, which makes the opinion analysis process error-prone. In this paper, we propose to employ bidirectional encoder representations transformers (BERT), and aspect-based sentiment analysis approaches in order to extract the relation between context dialogue sequence and response and determine whether or not the response is sarcastic. The best performing method of ours obtains an F1 score of 0.73 on the Twitter dataset and 0.734 over the Reddit dataset at the second workshop on figurative language processing Shared Task 2020.
Sarcasm analysis in user conversion text is automatic detection of any irony, insult, hurting, painful, caustic, humour, vulgarity that degrades an individual. It is helpful in the field of sentimental analysis and cyberbullying. As an immense growth of social media, sarcasm analysis helps to avoid insult, hurts and humour to affect someone. In this paper, we present traditional machine learning approaches, deep learning approach (LSTM -RNN) and BERT (Bidirectional Encoder Representations from Transformers) for identifying sarcasm. We have used the approaches to build the model, to identify and categorize how much conversion context or response is needed for sarcasm detection and evaluated on the two social media forums that is twitter conversation dataset and reddit conversion dataset. We compare the performance based on the approaches and obtained the best F1 scores as 0.722, 0.679 for the twitter forums and reddit forums respectively.
Social media platforms and discussion forums such as Reddit, Twitter, etc. are filled with figurative languages. Sarcasm is one such category of figurative language whose presence in a conversation makes language understanding a challenging task. In this paper, we present a deep neural architecture for sarcasm detection. We investigate various pre-trained language representation models (PLRMs) like BERT, RoBERTa, etc. and fine-tune it on the Twitter dataset. We experiment with a variety of PLRMs either on the twitter utterance in isolation or utilizing the contextual information along with the utterance. Our findings indicate that by taking into consideration the previous three most recent utterances, the model is more accurately able to classify a conversation as being sarcastic or not. Our best performing ensemble model achieves an overall F1 score of 0.790, which ranks us second on the leaderboard of the Sarcasm Shared Task 2020.
In this paper, we present the results obtained by BERT, BiLSTM and SVM classifiers on the shared task on Sarcasm Detection held as part of The Second Workshop on Figurative Language Processing. The shared task required the use of conversational context to detect sarcasm. We experimented by varying the amount of context used along with the response (response is the text to be classified). The amount of context used includes (i) zero context, (ii) last one, two or three utterances, and (iii) all utterances. It was found that including the last utterance in the dialogue along with the response improved the performance of the classifier for the Twitter data set. On the other hand, the best performance for the Reddit data set was obtained when using only the response without any contextual information. The BERT classifier obtained F-score of 0.743 and 0.658 for the Twitter and Reddit data set respectively.
Sarcasm Detection with Context, a shared task of Second Workshop on Figurative Language Processing (co-located with ACL 2020), is study of effect of context on Sarcasm detection in conversations of Social media. We present different techniques and models, mostly based on transformer for Sarcasm Detection with Context. We extended latest pre-trained transformers like BERT, RoBERTa, spanBERT on different task objectives like single sentence classification, sentence pair classification, etc. to understand role of conversation context for sarcasm detection on Twitter conversations and conversation threads from Reddit. We also present our own architecture consisting of LSTM and Transformers to achieve the objective.
Online discussion platforms are often flooded with opinions from users across the world on a variety of topics. Many such posts, comments, or utterances are often sarcastic in nature, i.e., the actual intent is hidden in the sentence and is different from its literal meaning, making the detection of such utterances challenging without additional context. In this paper, we propose a novel deep learning-based approach to detect whether an utterance is sarcastic or non-sarcastic by utilizing the given contexts ina hierarchical manner. We have used datasets from two online discussion platforms - Twitter and Reddit1for our experiments. Experimental and error analysis shows that the hierarchical models can make full use of history to obtain a better representation of contexts and thus, in turn, can outperform their sequential counterparts.
Sarcasm detection, regarded as one of the sub-problems of sentiment analysis, is a very typical task because the introduction of sarcastic words can flip the sentiment of the sentence itself. To date, many research works revolve around detecting sarcasm in one single sentence and there is very limited research to detect sarcasm resulting from multiple sentences. Current models used Long Short Term Memory (LSTM) variants with or without attention to detect sarcasm in conversations. We showed that the models using state-of-the-art Bidirectional Encoder Representations from Transformers (BERT), to capture syntactic and semantic information across conversation sentences, performed better than the current models. Based on the data analysis, we estimated that the number of sentences in the conversation that can contribute to the sarcasm and the results agrees to this estimation. We also perform a comparative study of our different versions of BERT-based model with other variants of LSTM model and XLNet (both using the estimated number of conversation sentences) and find out that BERT-based models outperformed them.
This paper reports a linguistically-enriched method of detecting token-level metaphors for the second shared task on Metaphor Detection. We participate in all four phases of competition with both datasets, i.e. Verbs and AllPOS on the VUA and the TOFEL datasets. We use the modality exclusivity and embodiment norms for constructing a conceptual representation of the nodes and the context. Our system obtains an F-score of 0.652 for the VUA Verbs track, which is 5% higher than the strong baselines. The experimental results across models and datasets indicate the salient contribution of using modality exclusivity and modality shift information for predicting metaphoricity.
In our daily life, metaphor is a common way of expression. To understand the meaning of a metaphor, we should recognize the metaphor words which play important roles. In the metaphor detection task, we design a sequence labeling model based on ALBERT-LSTM-softmax. By applying this model, we carry out a lot of experiments and compare the experimental results with different processing methods, such as with different input sentences and tokens, or the methods with CRF and softmax. Then, some tricks are adopted to improve the experimental results. Finally, our model achieves a 0.707 F1-score for the all POS subtask and a 0.728 F1-score for the verb subtask on the TOEFL dataset.
Recent work on automatic sequential metaphor detection has involved recurrent neural networks initialized with different pre-trained word embeddings and which are sometimes combined with hand engineered features. To capture lexical and orthographic information automatically, in this paper we propose to add character based word representation. Also, to contrast the difference between literal and contextual meaning, we utilize a similarity network. We explore these components via two different architectures - a BiLSTM model and a Transformer Encoder model similar to BERT to perform metaphor identification. We participate in the Second Shared Task on Metaphor Detection on both the VUA and TOFEL datasets with the above models. The experimental results demonstrate the effectiveness of our method as it outperforms all the systems which participated in the previous shared task.
This work explores the differences and similarities between neural image classifiers’ mis-categorisations and visually grounded metaphors - that we could conceive as intentional mis-categorisations. We discuss the possibility of using automatic image classifiers to approximate human metaphoric behaviours, and the limitations of such frame. We report two pilot experiments to study grounded metaphoricity. In the first we represent metaphors as a form of visual mis-categorisation. In the second we model metaphors as a more flexible, compositional operation in a continuous visual space generated from automatic classification systems.
This paper presents the first research aimed at recognizing euphemistic and dysphemistic phrases with natural language processing. Euphemisms soften references to topics that are sensitive, disagreeable, or taboo. Conversely, dysphemisms refer to sensitive topics in a harsh or rude way. For example, “passed away” and “departed” are euphemisms for death, while “croaked” and “six feet under” are dysphemisms for death. Our work explores the use of sentiment analysis to recognize euphemistic and dysphemistic language. First, we identify near-synonym phrases for three topics (firing, lying, and stealing) using a bootstrapping algorithm for semantic lexicon induction. Next, we classify phrases as euphemistic, dysphemistic, or neutral using lexical sentiment cues and contextual sentiment analysis. We introduce a new gold standard data set and present our experimental results for this task.
Metaphors are rhetorical use of words based on the conceptual mapping as opposed to their literal use. Metaphor detection, an important task in language understanding, aims to identify metaphors in word level from given sentences. We present IlliniMet, a system to automatically detect metaphorical words. Our model combines the strengths of the contextualized representation by the widely used RoBERTa model and the rich linguistic information from external resources such as WordNet. The proposed approach is shown to outperform strong baselines on a benchmark dataset. Our best model achieves F1 scores of 73.0% on VUA ALLPOS, 77.1% on VUA VERB, 70.3% on TOEFL ALLPOS and 71.9% on TOEFL VERB.
Metaphor processing and understanding has attracted the attention of many researchers recently with an increasing number of computational approaches. A common factor among these approaches is utilising existing benchmark datasets for evaluation and comparisons. The availability, quality and size of the annotated data are among the main difficulties facing the growing research area of metaphor processing. The majority of current approaches pertaining to metaphor processing concentrate on word-level processing due to data availability. On the other hand, approaches that process metaphors on the relation-level ignore the context where the metaphoric expression. This is due to the nature and format of the available data. Word-level annotation is poorly grounded theoretically and is harder to use in downstream tasks such as metaphor interpretation. The conversion from word-level to relation-level annotation is non-trivial. In this work, we attempt to fill this research gap by adapting three benchmark datasets, namely the VU Amsterdam metaphor corpus, the TroFi dataset and the TSV dataset, to suit relation-level metaphor identification. We publish the adapted datasets to facilitate future research in relation-level metaphor processing.
In this paper we describe computational ethnography study to demonstrate how machine learning techniques can be utilized to exploit bias resident in language data produced by communities with online presence. Specifically, we leverage the use of figurative language (i.e., the choice of metaphors) in online text (e.g., news media, blogs) produced by distinct communities to obtain models of community worldviews that can be shown to be distinctly biased and thus different from other communities’ models. We automatically construct metaphor-based community models for two distinct scenarios: gun rights and marriage equality. We then conduct a series of experiments to validate the hypothesis that the metaphors found in each community’s online language convey the bias in the community’s worldview.
This paper contains a preliminary corpus study of oxymorons, a figure of speech so far under-investigated in NLP-oriented research. The study resulted in a list of 376 oxymorons, identified by extracting a set of antonymous pairs (under various configurations) from corpora of written Italian and by manually checking the results. A complementary method is also envisaged for discovering contextual oxymorons, which are highly relevant for the detection of humor, irony and sarcasm.
Understanding and identifying humor has been increasingly popular, as seen by the number of datasets created to study humor. However, one area of humor research, humor generation, has remained a difficult task, with machine generated jokes failing to match human-created humor. As many humor prediction datasets claim to aid in generative tasks, we examine whether these claims are true. We focus our experiments on the most popular dataset, included in the 2020 SemEval’s Task 7, and teach our model to take normal text and “translate” it into humorous text. We evaluate our model compared to humorous human generated headlines, finding that our model is preferred equally in A/B testing with the human edited versions, a strong success for humor generation, and is preferred over an intelligent random baseline 72% of the time. We also show that our model is assumed to be human written comparable with that of the human edited headlines and is significantly better than random, indicating that this dataset does indeed provide potential for future humor generation systems.
This paper describes systems submitted to the Metaphor Shared Task at the Second Workshop on Figurative Language Processing. In this submission, we replicate the evaluation of the Bi-LSTM model introduced by Gao et al.(2018) on the VUA corpus in a new setting: TOEFL essays written by non-native English speakers. Our results show that Bi-LSTM models outperform feature-rich linear models on this challenging task, which is consistent with prior findings on the VUA dataset. However, the Bi-LSTM models lag behind the best performing systems in the shared task.
In this paper we present a novel resource-inexpensive architecture for metaphor detection based on a residual bidirectional long short-term memory and conditional random fields. Current approaches on this task rely on deep neural networks to identify metaphorical words, using additional linguistic features or word embeddings. We evaluate our proposed approach using different model configurations that combine embeddings, part of speech tags, and semantically disambiguated synonym sets. This evaluation process was performed using the training and testing partitions of the VU Amsterdam Metaphor Corpus. We use this method of evaluation as reference to compare the results with other current neural approaches for this task that implement similar neural architectures and features, and that were evaluated using this corpus. Results show that our system achieves competitive results with a simpler architecture compared to previous approaches.
The idea that a shift in concreteness within a sentence indicates the presence of a metaphor has been around for a while. However, recent methods of detecting metaphor that have relied on deep neural models have ignored concreteness and related psycholinguistic information. We hypothesis that this information is not available to these models and that their addition will boost the performance of these models in detecting metaphor. We test this hypothesis on the Metaphor Detection Shared Task 2020 and find that the addition of concreteness information does in fact boost deep neural models. We also run tests on data from a previous shared task and show similar results.
Supervised disambiguation of verbal idioms (VID) poses special demands on the quality and quantity of the annotated data used for learning and evaluation. In this paper, we present a new VID corpus for German and perform a series of VID disambiguation experiments on it. Our best classifier, based on a neural architecture, yields an error reduction across VIDs of 57% in terms of accuracy compared to a simple majority baseline.
We report the results of our system on the Metaphor Detection Shared Task at the Second Workshop on Figurative Language Processing 2020. Our model is an ensemble, utilising contextualised and static distributional semantic representations, along with word-type concreteness ratings. Using these features, it predicts word metaphoricity with a deep multi-layer perceptron. We are able to best the state-of-the-art from the 2018 Shared Task by an average of 8.0% F1, and finish fourth in both sub-tasks in which we participate.
Existing approaches to metaphor processing typically rely on local features, such as immediate lexico-syntactic contexts or information within a given sentence. However, a large body of corpus-linguistic research suggests that situational information and broader discourse properties influence metaphor production and comprehension. In this paper, we present the first neural metaphor processing architecture that models a broader discourse through the use of attention mechanisms. Our models advance the state of the art on the all POS track of the 2018 VU Amsterdam metaphor identification task. The inclusion of discourse-level information yields further significant improvements.
This paper describes the ETS entry to the 2020 Metaphor Detection shared task. Our contribution consists of a sequence of experiments using BERT, starting with a baseline, strengthening it by spell-correcting the TOEFL corpus, followed by a multi-task learning setting, where one of the tasks is the token-level metaphor classification as per the shared task, while the other is meant to provide additional training that we hypothesized to be relevant to the main task. In one case, out-of-domain data manually annotated for metaphor is used for the auxiliary task; in the other case, in-domain data automatically annotated for idioms is used for the auxiliary task. Both multi-task experiments yield promising results.
In this paper we present our results from the Second Shared Task on Metaphor Detection, hosted by the Second Workshop on Figurative Language Processing. We use an ensemble of RNN models with bidirectional LSTMs and bidirectional attention mechanisms. Some of the models were trained on all parts of speech. Each of the other models was trained on one of four categories for parts of speech: “nouns”, “verbs”, “adverbs/adjectives”, or “other”. The models were combined into voting pools and the voting pools were combined using the logical “OR” operator.
The detection of metaphors can provide valuable information about a given text and is crucial to sentiment analysis and machine translation. In this paper, we outline the techniques for word-level metaphor detection used in our submission to the Second Shared Task on Metaphor Detection. We propose using both BERT and XLNet language models to create contextualized embeddings and a bi-directional LSTM to identify whether a given word is a metaphor. Our best model achieved F1-scores of 68.0% on VUA AllPOS, 73.0% on VUA Verbs, 66.9% on TOEFL AllPOS, and 69.7% on TOEFL Verbs, placing 7th, 6th, 5th, and 5th respectively. In addition, we outline another potential approach with a KNN-LSTM ensemble model that we did not have enough time to implement given the deadline for the competition. We show that a KNN classifier provides a similar F1-score on a validation set as the LSTM and yields different information on metaphors.
This paper describes the adaptation and application of a neural network system for the automatic detection of metaphors. The LSTM BiRNN system participated in the shared task of metaphor identification that was part of the Second Workshop of Figurative Language Processing (FigLang2020) held at the Annual Conference of the Association for Computational Linguistics (ACL2020). The particular focus of our approach is on the potential influence that the metadata given in the ETS Corpus of Non-Native Written English might have on the automatic detection of metaphors in this dataset. The article first discusses the annotated ETS learner data, highlighting some of its peculiarities and inherent biases of metaphor use. A series of evaluations follow in order to test whether specific metadata influence the system performance in the task of automatic metaphor identification. The system is available under the APLv2 open-source license.
We present an ensemble approach for the detection of sarcasm in Reddit and Twitter responses in the context of The Second Workshop on Figurative Language Processing held in conjunction with ACL 2020. The ensemble is trained on the predicted sarcasm probabilities of four component models and on additional features, such as the sentiment of the comment, its length, and source (Reddit or Twitter) in order to learn which of the component models is the most reliable for which input. The component models consist of an LSTM with hashtag and emoji representations; a CNN-LSTM with casing, stop word, punctuation, and sentiment representations; an MLP based on Infersent embeddings; and an SVM trained on stylometric and emotion-based features. All component models use the two conversational turns preceding the response as context, except for the SVM, which only uses features extracted from the response. The ensemble itself consists of an adaboost classifier with the decision tree algorithm as base estimator and yields F1-scores of 67% and 74% on the Reddit and Twitter test data, respectively.
Understanding tone in Twitter posts will be increasingly important as more and more communication moves online. One of the most difficult, yet important tones to detect is sarcasm. In the past, LSTM and transformer architecture models have been used to tackle this problem. We attempt to expand upon this research, implementing LSTM, GRU, and transformer models, and exploring new methods to classify sarcasm in Twitter posts. Among these, the most successful were transformer models, most notably BERT. While we attempted a few other models described in this paper, our most successful model was an ensemble of transformer models including BERT, RoBERTa, XLNet, RoBERTa-large, and ALBERT. This research was performed in conjunction with the sarcasm detection shared task section in the Second Workshop on Figurative Language Processing, co-located with ACL 2020.
We present a transformer-based sarcasm detection model that accounts for the context from the entire conversation thread for more robust predictions. Our model uses deep transformer layers to perform multi-head attentions among the target utterance and the relevant context in the thread. The context-aware models are evaluated on two datasets from social media, Twitter and Reddit, and show 3.1% and 7.0% improvements over their baselines. Our best models give the F1-scores of 79.0% and 75.0% for the Twitter and Reddit datasets respectively, becoming one of the highest performing systems among 36 participants in this shared task.
Framenets as an incarnation of frame semantics have been set up to deal with lexicographic issues (cf. Fillmore and Baker 2010, among others). They are thus concerned with lexical units (LUs) and the conceptual structure which categorizes these together. These lexically-evoked frames, however, do not reflect pragmatic properties of constructions (LUs and other types of constructions), such as expressing illocutions or being considered polite or very informal. From the viewpoint of a multilingual annotation effort, the Global FrameNet Shared Annotation Task, we discuss two phenomena, greetings and tag questions, which highlight the necessity both to investigate the role between construction and frame annotation on the one hand and to develop pragmatic frames describing social interactions which are not explicitly lexicalized.
This paper reports on an effort to search for corresponding constructions in English and Japanese in a TED Talk parallel corpus, using frames-and-constructions analysis (Ohara, 2019; Ohara and Okubo, 2020; cf. Czulo, 2013, 2017). The purpose of the paper is two-fold: (1) to demonstrate the validity of frames-and-constructions analysis to search for corresponding constructions in typologically unrelated languages; and (2) to assess whether the “Do schools kill creativity?” TED Talk parallel corpus, annotated in various languages for Multilingual FrameNet, is a good starting place for building a multilingual constructicon. The analysis showed that similar to our previous findings involving texts in a Japanese to English bilingual children’s book, the TED Talk bilingual transcripts include pairs of constructions that share similar pragmatic functions. While the TED Talk parallel corpus constitutes a good resource for frame semantic annotation in multiple languages, it may not be the ideal place to start aligning constructions among typologically unrelated languages. Finally, this work shows that the proposed method, which focuses on heads of sentences, seems valid for searching for corresponding constructions in transcripts of spoken data, as well as in written data of typologically-unrelated languages.
In this paper, we introduce the task of using FrameNet to link structured information about real-world events to the conceptual frames used in texts describing these events. We show that frames made relevant by the knowledge of the real-world event can be captured by complementing standard lexicon-driven FrameNet annotations with frame annotations derived through pragmatic inference. We propose a two-layered annotation scheme with a ‘strict’ FrameNet-compatible lexical layer and a ‘loose’ layer capturing frames that are inferred from referential data.
Multimodal aspects of human communication are key in several applications of Natural Language Processing, such as Machine Translation and Natural Language Generation. Despite recent advances in integrating multimodality into Computational Linguistics, the merge between NLP and Computer Vision techniques is still timid, especially when it comes to providing fine-grained accounts for meaning construction. This paper reports on research aiming to determine appropriate methodology and develop a computational tool to annotate multimodal corpora according to a principled structured semantic representation of events, relations and entities: FrameNet. Taking a Brazilian television travel show as corpus, a pilot study was conducted to annotate the frames that are evoked by the audio and the ones that are evoked by visual elements. We also implemented a Multimodal Annotation tool which allows annotators to choose frames and locate frame elements both in the text and in the images, while keeping track of the time span in which those elements are active in each modality. Results suggest that adding a multimodal domain to the linguistic layer of annotation and analysis contributes both to enrich the kind of information that can be tagged in a corpus, and to enhance FrameNet as a model of linguistic cognition.
We introduce an annotation tool whose purpose is to gain insights into variation of framing by combining FrameNet annotation with referential annotation. English FrameNet enables researchers to study variation in framing at the conceptual level as well through its packaging in language. We enrich FrameNet annotations in two ways. First, we introduce the referential aspect. Secondly, we annotate on complete texts to encode connections between mentions. As a result, we can analyze the variation of framing for one particular event across multiple mentions and (cross-lingual) documents. We can examine how an event is framed over time and how core frame elements are expressed throughout a complete text. The data model starts with a representation of an event type. Each event type has many incidents linked to it, and each incident has several reference texts describing it as well as structured data about the incident. The user can apply two types of annotations: 1) mappings from expressions to frames and frame elements, 2) reference relations from mentions to events and participants of the structured data.
This paper presents an approach to project FrameNet annotations into other languages using attention-based neural machine translation (NMT) models. The idea is to use an NMT encoder-decoder attention matrix to propose a word-to-word correspondence between the source and the target language. We combine this word alignment along with a set of simple rules to securely project the FrameNet annotations into the target language. We successfully implemented, evaluated and analyzed this technique on the English-to-French configuration. First, we analyze the obtained FrameNet lexicon qualitatively. Then, we use existing French FrameNet corpora to assert the quality of the translation. Finally, we trained a BERT-based FrameNet parser using the projected annotations and compared it to a BERT baseline. Results show substantial improvements in the French language, giving evidence to support that our approach could help to propagate FrameNet data-set on other languages.
Large coverage lexical resources that bear deep linguistic information have always been considered useful for many natural language processing (NLP) applications including Machine Translation (MT). In this respect, Frame-based resources have been developed for many languages following Frame Semantics and the Berkeley FrameNet project. However, to a great extent, all those efforts have been kept fragmented. Consequentially, the Global FrameNet initiative has been conceived of as a joint effort to bring together FrameNets in different languages. The proposed paper is aimed at describing ongoing work towards developing the Greek (EL) counterpart of the Global FrameNet and our efforts to contribute to the Shared Annotation Task. In the paper, we will elaborate on the annotation methodology employed, the current status and progress made so far, as well as the problems raised during annotation.
This paper presents the first investigation on using semantic frames to assess text difficulty. Based on Mandarin VerbNet, a verbal semantic database that adopts a frame-based approach, we examine usage patterns of ten verbs in a corpus of graded Chinese texts. We identify a number of characteristics in texts at advanced grades: more frequent use of non-core frame elements; more frequent omission of some core frame elements; increased preference for noun phrases rather than clauses as verb arguments; and more frequent metaphoric usage. These characteristics can potentially be useful for automatic prediction of text readability.
We propose an approach for generating an accurate and consistent PropBank-annotated corpus, given a FrameNet-annotated corpus which has an underlying dependency annotation layer, namely, a parallel Universal Dependencies (UD) treebank. The PropBank annotation layer of such a multi-layer corpus can be semi-automatically derived from the existing FrameNet and UD annotation layers, by providing a mapping configuration from lexical units in [a non-English language] FrameNet to [English language] PropBank predicates, and a mapping configuration from FrameNet frame elements to PropBank semantic arguments for the given pair of a FrameNet frame and a PropBank predicate. The latter mapping generally depends on the underlying UD syntactic relations. To demonstrate our approach, we use Latvian FrameNet, annotated on top of Latvian UD Treebank, for generating Latvian PropBank in compliance with the Universal Propositions approach.
The Emirati Arabic FrameNet (EAFN) project aims to initiate a FrameNet for Emirati Arabic, utilizing the Emirati Arabic Corpus. The goal is to create a resource comparable to the initial stages of the Berkeley FrameNet. The project is divided into manual and automatic tracks, based on the predominant techniques being used to collect frames in each track. Work on the EAFN is progressing, and we here report on initial results for annotations and evaluation. The EAFN project aims to provide a general semantic resource for the Arabic language, sure to be of interest to researchers from general linguistics to natural language processing. As we report here, the EAFN is well on target for the first release of data in the coming year.
The FrameNet (FN) project at the International Computer Science Institute in Berkeley (ICSI), which documents the core vocabulary of contemporary English, was the first lexical resource based on Fillmore’s theory of Frame Semantics. Berkeley FrameNet has inspired related projects in roughly a dozen other languages, which have evolved somewhat independently; the current Multilingual FrameNet project (MLFN) is an attempt to find alignments between all of them. The alignment problem is complicated by the fact that these projects have adhered to the Berkeley FrameNet model to varying degrees, and they were also founded at different times, when different versions of the Berkeley FrameNet data were available. We describe several new methods for finding relations of similarity between semantic frames across languages. We will demonstrate ViToXF, a new tool which provides interactive visualizations of these cross-lingual relations, between frames, lexical units, and frame elements, based on resources such as multilingual dictionaries and on shared distributional vector spaces, making clear the strengths and weaknesses of different alignment methods.
The methodology developed within the FrameNet project is being used to compile resources in an increasing number of specialized fields of knowledge. The methodology along with the theoretical principles on which it is based, i.e. Frame Semantics, are especially appealing as they allow domain-specific resources to account for the conceptual background of specialized knowledge and to explain the linguistic properties of terms against this background. This paper presents a methodology for building a multilingual resource that accounts for terms of the environment. After listing some lexical and conceptual differences that need to be managed in such a resource, we explain how the FrameNet methodology is adapted for describing terms in different languages. We first applied our methodology to French and then extended it to English. Extensions to Spanish, Portuguese and Chinese were made more recently. Up to now, we have defined 190 frames: 112 frames are new; 38 are used as such; and 40 are slightly different (a different number of obligatory participants; a significant alternation, etc.) when compared to Berkeley FrameNet.
A weak point of rule-based sentiment analysis systems is that the underlying sentiment lexicons are often not adapted to the domain of the text we want to analyze. We created a game-specific sentiment lexicon for video game Skyrim based on the E-ANEW word list and a dataset of Skyrim’s in-game documents. We calculated sentiment ratings for NPC dialogue using both our lexicon and E-ANEW and compared the resulting sentiment ratings to those of human raters. Both lexicons perform comparably well on our evaluation dialogues, but the game-specific extension performs slightly better on the dominance dimension for dialogue segments and the arousal dimension for full dialogues. To our knowledge, this is the first time that a sentiment analysis lexicon has been adapted to the video game domain.
The ESP Game (also known as the Google Image Labeler) demonstrated how the crowd could perform a task that is straightforward for humans but challenging for computers – providing labels for images. The game facilitated the task of basic image labeling; however, the labels generated were non-specific and limited the ability to distinguish similar images from one another, limiting its ability in search tasks, annotating images for the visually impaired, and training computer vision machine algorithms. In this paper, we describe ClueMeIn, an entertaining web-based game with a purpose that generates more detailed image labels than the ESP Game. We conduct experiments to generate specific image labels, show how the results can lead to improvements in the accuracy of image searches over image labels generated by the ESP Game when using the same public dataset.
Errors commonly exist in machine-generated documents and publication materials; however, some correction algorithms do not perform well for complex errors and it is costly to employ humans to do the task. To solve the problem, a prototype computer game called Cipher was developed that encourages people to identify errors in text. Gamification is achieved by introducing the idea of steganography as the entertaining game element. People play the game for entertainment while they make valuable annotations to locate text errors. The prototype was tested by 35 players in a evaluation experiment, creating 4,764 annotations. After filtering the data, the system detected manually introduced text errors and also genuine errors in the texts that were not noticed when they were introduced into the game.
GWAP design might have a tremendous effect on its popularity of course but also on the quality of the data collected. In this paper, a comparison is undertaken between two GWAPs for building term association lists, namely JeuxDeMots and Quicky Goose. After comparing both game designs, the Cohen kappa of associative lists in various configurations is computed in order to assess likeness and differences of the data they provide.
In this paper, we describe a Telegram bot, Mago della Ghigliottina (Ghigliottina Wizard), able to solve La Ghigliottina game (The Guillotine), the final game of the Italian TV quiz show L’Eredità. Our system relies on linguistic resources and artificial intelligence and achieves better results than human players (and competitors of L’Eredità too). In addition to solving a game, Mago della Ghigliottina can also generate new game instances and challenge the users to match the solution.
Gamification has been applied to many linguistic annotation tasks, as an alternative to crowdsourcing platforms to collect annotated data in an inexpensive way. However, we think that still much has to be explored. Games with a Purpose (GWAPs) tend to lack important elements that we commonly see in commercial games, such as 2D and 3D worlds or a story. Making GWAPs more similar to full-fledged video games in order to involve users more easily and increase dissemination is a demanding yet interesting ground to explore. In this paper we present a 3D role-playing game for abusive language annotation that is currently under development.
In this paper we present a new method for collecting naturally generated dialogue data for a low resourced language, (specifically here—Uyghur). We plan to build a games with a purpose (GWAPs) to encourage native speakers to actively contribute dialogue data to our research project. Since we aim to characterize the response space of queries in Uyghur, we design various scenarios for conversations that yield to questions being posed and responded to. We will implement the GWAP with the RPG Maker MV Game Engine, and will integrate the chatroom system in the game with the Dialogue Experimental Toolkit (DiET). DiET will help us improve the data collection process, and most importantly, make us have some control over the interactions among the participants.
In this paper, we present the ongoing development of CALLIG – a web system that uses improvisation games in Computer Assisted Language Learning (CALL). Improvisation games are structured activities with built-in constraints where improvisers are asked to generate a lot of different ideas and weave a diverse range of elements into a sensible narrative spontaneously. This paper discusses how computer-based language games can be created combining improvisation elements and language technology. In contrast with traditional language exercises, improvisational language games are open and unpredictable. CALLIG encourages spontaneity and witty language use. It also provides opportunities for collecting useful data for many NLP applications.
Although the roguelike video game genre has a large community of fans (both players and developers) and the graphic aspect of these games is usually given little relevance (ASCII-based graphics are not rare even today), their accessibility for blind players and other visually-impaired users remains a pending issue. In this document, we describe an initiative for the development of roguelikes adapted to visually-impaired players by using Natural Language Processing techniques, together with the first completed games resulting from it. These games were developed as Bachelor’s and Master’s theses. Our approach consists in integrating a multilingual module that, apart from the classic ASCII-based graphical interface, automatically generates text descriptions of what is happening within the game. The visually-impaired user can then read such descriptions by means of a screen reader. In these projects we seek expressivity and variety in the descriptions, so we can offer the users a fun roguelike experience that does not sacrifice any of the key characteristics that define the genre. Moreover, we intend to make these projects easy to extend to other languages, thus avoiding costly and complex solutions. KEYWORDS: Natural Language Generation, roguelikes, visually-impaired users
Increasing efforts are put into gamification of experimentation software in psychology and educational applications and the development of serious games. Computer-based experiments with game-like features have been developed previously for research on cognitive skills, cognitive processing speed, working memory, attention, learning, problem solving, group behavior and other phenomena. It has been argued that computer game experiments are superior to traditional computerized experimentation methods in laboratory tasks in that they represent holistic, meaningful, and natural human activity. We present a novel experimental framework for forced choice categorization tasks or speech perception studies in the form of a computer game, based on the Unity Engine – the Gamified Discrimination Experiments engine (GDX). The setting is that of a first person shooter game with the narrative background of an alien invasion on earth. We demonstrate the utility of our game as a research tool with an application focusing on attention to fine phonetic detail in natural speech perception. The game-based framework is additionally compared against a traditional experimental setup in an auditory discrimination task. Applications of this novel game-based framework are multifarious within studies on all aspects of spoken language perception.
As the uses of Games-With-A-Purpose (GWAPs) broadens, the systems that incorporate its usages have expanded in complexity. The types of annotations required within the NLP paradigm set such an example, where tasks can involve varying complexity of annotations. Assigning more complex tasks to more skilled players through a progression mechanism can achieve higher accuracy in the collected data while acting as a motivating factor that rewards the more skilled players. In this paper, we present the progression technique implemented in Wormingo , an NLP GWAP that currently includes two layers of task complexity. For the experiment, we have implemented four different progression scenarios on 192 players and compared the accuracy and engagement achieved with each scenario.
While playing the communication game “Are You a Werewolf”, a player always guesses other players’ roles through discussions, based on his own role and other players’ crucial utterances. The underlying goal of this paper is to construct an agent that can analyze the participating players’ utterances and play the werewolf game as if it is a human. For a step of this underlying goal, this paper studies how to accumulate werewolf game log data annotated with identification of players revealing oneselves as seer/medium, the acts of the divination and the medium and declaring the results of the divination and the medium. In this paper, we divide the whole task into four sub tasks and apply CNN/SVM classifiers to each sub task and evaluate their performance.
The OntoLex vocabulary enjoys increasing popularity as a means of publishing lexical resources with RDF and as Linked Data. The recent publication of a new OntoLex module for lexicography, lexicog, reflects its increasing importance for digital lexicography. However, not all aspects of digital lexicography have been covered to the same extent. In particular, supplementary information drawn from corpora such as frequency information, links to attestations, and collocation data were considered to be beyond the scope of lexicog. Therefore, the OntoLex community has put forward the proposal for a novel module for frequency, attestation and corpus information (FrAC), that not only covers the requirements of digital lexicography, but also accommodates essential data structures for lexical information in natural language processing. This paper introduces the current state of the OntoLex-FrAC vocabulary, describes its structure, some selected use cases, elementary concepts and fundamental definitions, with a focus on frequency and attestations.
This paper reports on an extended version of a synonym verb class lexicon, newly called SynSemClass (formerly CzEngClass). This lexicon stores cross-lingual semantically similar verb senses in synonym classes extracted from a richly annotated parallel corpus, the Prague Czech-English Dependency Treebank. When building the lexicon, we make use of predicate-argument relations (valency) and link them to semantic roles; in addition, each entry is linked to several external lexicons of more or less “semantic” nature, namely FrameNet, WordNet, VerbNet, OntoNotes and PropBank, and Czech VALLEX. The aim is to provide a linguistic resource that can be used to compare semantic roles and their syntactic properties and features across languages within and across synonym groups (classes, or ’synsets’), as well as gold standard data for automatic NLP experiments with such synonyms, such as synonym discovery, feature mapping, etc. However, perhaps the most important goal is to eventually build an event type ontology that can be referenced and used as a human-readable and human-understandable “database” for all types of events, processes and states. While the current paper describes primarily the content of the lexicon, we are also presenting a preliminary design of a format compatible with Linked Data, on which we are hoping to get feedback during discussions at the workshop. Once the resource (in whichever form) is applied to corpus annotation, deep analysis will be possible using such combined resources as training data.
In this paper we describe the process of inclusion of etymological information in a knowledge base of interoperable Latin linguistic resources developed in the context of the LiLa: Linking Latin project. Interoperability is obtained by applying the Linked Open Data principles. Particularly, an extensive collection of Latin lemmas is used to link the (distributed) resources. For the etymology, we rely on the Ontolex-lemon ontology and the lemonEty extension to model the information, while the source data are taken from a recent etymological dictionary of Latin. As a result, the collection of lemmas LiLa is built around now includes 1,465 Proto-Italic and 1,393 Proto-Indo-European reconstructed forms that are used to explain the history of 1,400 Latin words. We discuss the motivation, methodology and modeling strategies of the work, as well as its possible applications and potential future developments.
We present the ongoing work on an automatically generated dictionary describing Danish in the 16th century. A series of relevant dictionaries – from the period as well as more recent ones – are linked together at lemma level, and where possible, definitions or keywords are extracted and presented in the new dictionary.
This extended abstract presents on-going work consisting in interlinking and merging the Open Dutch WordNet and generic lexicographic resources for Dutch, focusing for now on the Dutch and English versions of Wiktionary and using the Algemeen Nederlands Woordenboek as a quality checking instance. As the Open Dutch WordNet is already equipped with a relevant number of complex lexical units, we are aiming at expanding it and proposing a new representational framework for the encoding of the interlinked and integrated data. The longer term goal of the work is to investigate if and on how senses can be restricted to particular morphological variations of Dutch lexical entries, and how to represent this information in a Linguistic Linked Open Data compliant format.
There are wordnets in many languages, many aligned with Princeton WordNet, some of which in a (semi-)automatic process, but we rarely see actual discussions on the role of false friends in this process. Having in mind known issues related to such words in language translation, and further motivated by false friend-related issues on the alignment of a Portuguese wordnet with Princeton Wordnet, we aim to widen this discussion, while suggesting preliminary ideas of how wordnets could benefit from this kind of research.
This paper describes the development and current state of Pinchah Kristang – an online dictionary for Kristang. Kristang is a critically endangered language of the Portuguese-Eurasian communities residing mainly in Malacca and Singapore. Pinchah Kristang has been a central tool to the revitalization efforts of Kristang in Singapore, and collates information from multiple sources, including existing dictionaries and wordlists, ongoing language documentation work, and new words that emerge regularly from relexification efforts by the community. This online dictionary is powered by the Princeton Wordnet and the Open Kristang Wordnet – a choice that brings both advantages and disadvantages. This paper will introduce the current version of this dictionary, motivate some of its design choices, and discuss possible future directions.
Our aim is to identify suitable sense representations for NLP in Danish. We investigate sense inventories that correlate with human interpretations of word meaning and ambiguity as typically described in dictionaries and wordnets and that are well reflected distributionally as expressed in word embeddings. To this end, we study a number of highly ambiguous Danish nouns and examine the effectiveness of sense representations constructed by combining vectors from a distributional model with the information from a wordnet. We establish representations based on centroids obtained from wordnet synests and example sentences as well as representations established via are tested in a word sense disambiguation task. We conclude that the more information extracted from the wordnet entries (example sentence, definition, semantic relations) the more successful the sense representation vector.
Bring’s thesaurus (Bring) is a Swedish counterpart of Roget, and its digitized version could make a valuable language resource for use in many and diverse natural language processing (NLP) applications. From the literature we know that Roget-style thesauruses and wordnets have complementary strengths in this context, so both kinds of lexical-semantic resource are good to have. However, Bring was published in 1930, and its lexical items are in the form of lemma–POS pairings. In order to be useful in our NLP systems, polysemous lexical items need to be disambiguated, and a large amount of modern vocabulary must be added in the proper places in Bring. The work presented here describes experiments aiming at automating these two tasks, at least in part, where we use the structure of an existing Swedish semantic lexicon – Saldo – both for disambiguation of ambiguous Bring entries and for addition of new entries to Bring.
An Excellency Research Project called “Terminology of olive oil and trade: China and other international markets” (P07-HUM-03041) was initiated under my management in 2008, financed by the Andalusian regional government, the Junta de Andalucía. The project, known as “OLIVATERM”, had two main objectives: on the one hand, to develop the first systematic multilingual terminological dictionary in the scientific and socio-economic area of the olive grove and olive oils in order to facilitate communication in the topic; on the other, to contribute to the expansion of the Andalusia’s domestic and international trade and the dissemination of its culture. The main outcome of the research was the Diccionario de términos del aceite de oliva (DTAO – Dictionary of olive oil terms) (Roldán Vendrell, Arco Libros: 2013). This dictionary is currently the main reference source for answering queries and responding to any doubts that might arise in the use of this terminology in the three reference languages (Spanish, English and Chinese). It has received unanimous acknowledgement from numerous specialists in the sphere of Terminology, including most especially Maria Teresa Cabré (UPF), Miguel Casas Gómez (UCA- Ibérica 27 (2014): 217-234), François Maniez (Université de Lyon), Maria Isabel Santamaría Pérez and Chelo Vargas Sierra (UA), Pamela Faber (UGR), Joaquín García Palacios (USAL), and Marie-Claude L’Homme (Université de Montréal). The DTAO is well-known in the academic area of Terminology, but has not reached many of the institutions and organizations (domestic and international), translators, journalists, communicators and olive oil sector professionals that could benefit from it in their professions, especially salespeople, who need (fortunately, with an ever greater frequency) information on terminology in the book’s target languages for their commercial transactions. That is why we are currently working on a multichannel technological solution that enables a greater and more efficient transfer to the business sector: the design and development of an adaptive website (responsive web design) that provides access to the information in any usage context. We believe that access must be afforded to this valuable reference information on a hand-held device that enables it to be looked up both on- and offline and so pre-empt situations in which it is impossible to connect to the internet. The web application’s database will therefore also feed a series of mobile applications that will be available for the main platforms (iOS, Android). This tool will represent real progress in the dynamic transfer of specialized knowledge in the field of olive growing and olive oil production. Apart from delivering universal and free access to this information, the web application will welcome user suggestions for including new terms, new information and new reference languages, making it a collaborative tool that is also fed by its own users. With this tool we hope to respond to society’s needs for multilingual communication in the area of olive oil and to help give a boost to economic activity in the olive sector. In this work, in parallel to the presentation of the adaptive website, we will present a lexical repertoire integrated by new terms and expressions coined in this field (in the three working languages) in the last years. These neologisms reflect the most relevant innovations occurred in the olive oil sector over the last decade and, therefore, they must be compiled, sorted, systematized, and made accessible to the users in the web application we intend to develop.
Thanks to new technologies, the elaboration of specialized bilingual dictionaries can be made faster and more standardized, offering not only a dictionary of equivalents, but also the representation of a conceptual field. Nevertheless, in view of these new tools and services, some of which are offered free of charge by European institutions, it is necessary to question the viability of their use by a lambda user and the previous knowledge required for such use, as well as the possible problems they may encounter. In our communication we show a series of possible difficulties, as well as a methodological proposal and some solutions, by presenting an extract of a French-Spanish bilingual dictionary for the domain of architecture. The extract in question is a sample of about 30 terms created with the Lexonomy dictionary editor (Měchura 2017).
This paper describes RACAI’s word sense alignment system, which participated in the Monolingual Word Sense Alignment shared task organized at GlobaLex 2020 workshop. We discuss the system architecture, some of the challenges that we faced as well as present our results on several of the languages available for the task.
In this paper we describe the system submitted to the ELEXIS Monolingual Word Sense Alignment Task. We test different systems,which are two types of LSTMs and a system based on a pretrained Bidirectional Encoder Representations from Transformers (BERT)model, to solve the task. LSTM models use fastText pre-trained word vectors features with different settings. For training the models,we did not combine external data with the dataset provided for the task. We select a sub-set of languages among the proposed ones,namely a set of Romance languages, i.e., Italian, Spanish, Portuguese, together with English and Dutch. The Siamese LSTM withattention and PoS tagging (LSTM-A) performed better than the other two systems, achieving a 5-Class Accuracy score of 0.844 in theOverall Results, ranking the first position among five teams.
This paper describes our system for monolingual sense alignment across dictionaries. The task of monolingual word sense alignment is presented as a task of predicting the relationship between two senses. We will present two solutions, one based on supervised machine learning, and the other based on pre-trained neural network language model, specifically BERT. Our models perform competitively for binary classification, reporting high scores for almost all languages. This paper presents our submission for the shared task on monolingual word sense alignment across dictionaries as part of the GLOBALEX 2020 – Linked Lexicography workshop at the 12th Language Resources and Evaluation Conference (LREC). Monolingual word sense alignment (MWSA) is the task of aligning word senses across re- sources in the same language. Lexical-semantic resources (LSR) such as dictionaries form valuable foundation of numerous natural language process- ing (NLP) tasks. Since they are created manually by ex- perts, dictionaries can be considered among the resources of highest quality and importance. However, the existing LSRs in machine readable form are small in scope or miss- ing altogether. Thus, it would be extremely beneficial if the existing lexical resources could be connected and ex- panded. Lexical resources display considerable variation in the number of word senses that lexicographers assign to a given entry in a dictionary. This is because the identification and differentiation of word senses is one of the harder tasks that lexicographers face. Hence, the task of combining dictio- naries from different sources is difficult, especially for the case of mapping the senses of entries, which often differ significantly in granularity and coverage. (Ahmadi et al., 2020) There are three different angles from which the problem of word sense alignment can be addressed: approaches based on the similarity of textual descriptions of word senses, ap- proaches based on structural properties of lexical-semantic resources, and a combination of both. (Matuschek, 2014) In this paper we focus on the similarity of textual de- scriptions. This is a common approach as the majority of previous work used some notion of similarity between senses, mostly gloss overlap or semantic relatedness based on glosses. This makes sense, as glosses are a prerequisite for humans to recognize the meaning of an encoded sense, and thus also an intuitive way of judging the similarity of senses. (Matuschek, 2014) The paper is structured as follows: we provide a brief overview of related work in Section 2, and a description of the corpus in Section 3. In Section 4 we explain all impor- tant aspects of our model implementation, while the results are presented in Section 5. Finally, we end the paper with the discussion in Section 6 and conclusion in Section 7.
In this paper, we present the NUIG system at the TIAD shard task. This system includes graph-based metrics calculated using novel algorithms, with an unsupervised document embedding tool called ONETA and an unsupervised multi-way neural machine translation method. The results are an improvement over our previous system and produce the highest precision among all systems in the task as well as very competitive F-Measure results. Incorporating features from other systems should be easy in the framework we describe in this paper, suggesting this could very easily be extended to an even stronger result.
This paper describes our contribution to the Third Shared Task on Translation Inference across Dictionaries (TIAD-2020). We describe an approach on translation inference based on symbolic methods, the propagation of concepts over a graph of interconnected dictionaries: Given a mapping from source language words to lexical concepts (e.g., synsets) as a seed, we use bilingual dictionaries to extrapolate a mapping of pivot and target language words to these lexical concepts. Translation inference is then performed by looking up the lexical concept(s) of a source language word and returning the target language word(s) for which these lexical concepts have the respective highest score. We present two instantiations of this system: One using WordNet synsets as concepts, and one using lexical entries (translations) as concepts. With a threshold of 0, the latter configuration is the second among participant systems in terms of F1 score. We also describe additional evaluation experiments on Apertium data, a comparison with an earlier approach based on embedding projection, and an approach for constrained projection that outperforms the TIAD-2020 vanilla system by a large margin.
This paper describes the participation of two different approaches in the 3rd Translation Inference Across Dictionaries (TIAD 2020) shared task. The aim of the task is to automatically generate new bilingual dictionaries from existing ones. To that end, we essayed two different types of techniques: based on graph exploration on the one hand and, on the other hand, based on cross-lingual word embeddings. The task evaluation results show that graph exploration is very effective, accomplishing relatively high precision and recall values in comparison with the other participating systems, while cross-lingual embeddings reaches high precision but smaller recall.
This paper describes four different strategies proposed to the TIAD 2020 Shared Task for automatic translation inference across dictionaries. The proposed strategies are based on the analysis of Apertium RDF graph, taking advantage of characteristics such as translation using multiple paths, synonyms and similarities between lexical entries from different lexicons and cardinality of possible translations through the graph. The four strategies were trained and validated on the Apertium RDF EN<->ES dictionary, showing promising results. Finally, the strategies, applied together, obtained an F-measure of 0.43 in the task of inferring the dictionaries proposed in the shared task, ranking thus third with respect to the other new systems presented to the TIAD 2020 Shared Task. No system presented to the shared task exceeded the baseline proposed by the TIAD organizers.
This paper discusses the current state of developing an ISO standard annotation scheme for quantification phenomena in natural language, as part of the ISO Semantic Annotation Framework (ISO 24617). A proposed approach that combines ideas from the theory of generalised quantifiers and from neo-Davidsonian event semantics was adopted by the ISO organisation in 2019 as a starting point for developing such an annotation scheme. * This scheme consists of (1) a conceptual ‘metamodel’ that visualises the types of entities, functions and relations that go into annotations of quantification; (2) an abstract syntax which defines ‘annotation structures’ as triples and other set-theoretic constructs; (3) an XML-based representation of annotation structures (‘concrete syntax’); and (4) a compositional semantics of annotation structures. The latter three components together define the interpreted markup language QuantML. The focus in this paper is on the structuring of the semantic information needed to characterise quantification in natural language and the representation of these structures in QuantML.
ISO-TimeML is an international standard for multilingual event annotation, detection, categorization and linking. In this paper, we present the Hindi TimeBank, an ISO-TimeML annotated reference corpus for the detection and classification of events, states and time expressions, and the links between them. Based on contemporary developments in Hindi event recognition, we propose language independent and language-specific deviations from the ISO-TimeML guidelines, but preserve the schema. These deviations include the inclusion of annotator confidence, and an independent mechanism of identifying and annotating states such as copulars and existentials) With this paper, we present an open-source corpus, the Hindi TimeBank. The Hindi TimeBank is a 1,000 article dataset, with over 25,000 events, 3,500 states and 2,000 time expressions. We analyze the dataset in detail and provide a class-wise distribution of events, states and time expressions. Our guidelines and dataset are backed by high average inter-annotator agreement scores.
The paper presents an annotation schema with the following characteristics: it is formally compact; it systematically and compositionally expands into fullfledged analytic representations, exploiting simple algorithms of typed feature structures; its representation of various dimensions of semantic content is systematically integrated with morpho-syntactic and lexical representation; it is integrated with a ‘deep’ parsing grammar. Its compactness allows for efficient handling of large amounts of structures and data, and it is interoperable in covering multiple aspects of grammar and meaning. The code and its analytic expansions represent a cross-linguistically wide range of phenomena of languages and language structures. This paper presents its syntactic-semantic interoperability first from a theoretical point of view and then as applied in linguistic description.
People’s visual perception is very pronounced and therefore it is usually no problem for them to describe the space around them in words. Conversely, people also have no problems imagining a concept of a described space. In recent years many efforts have been made to develop a linguistic concept for spatial and spatial-temporal relations. However, the systems have not really caught on so far, which in our opinion is due to the complex models on which they are based and the lack of available training data and automated taggers. In this paper we describe a project to support spatial annotation, which could facilitate annotation by its many functions, but also enrich it with many more information. This is to be achieved by an extension by means of a VR environment, with which spatial relations can be better visualized and connected with real objects. And we want to use the available data to develop a new state-of-the-art tagger and thus lay the foundation for future systems such as improved text understanding for Text2Scene.
This paper proposes a semantics ABS for the model-theoretic interpretation of annotation structures. It provides a language ABSr, that represents semantic forms in a (possibly 𝜆-free) type-theoretic first-order logic. For semantic compositionality, the representation language introduces two operators ⊕ and ⊘ with subtypes for the conjunctive or distributive composition of semantic forms. ABS also introduces a small set of logical predicates to represent semantic forms in a simplified format. The use of ABSr is illustrated with some annotation structures that conform to ISO 24617 standards on semantic annotation such as ISO-TimeML and ISO-Space.
This short research paper presents the results of a corpus-based metonymy annotation exercise on a sample of 101 Croatian verb entries – corresponding to 457 patters and over 20,000 corpus lines – taken from CROATPAS (Marini & Ježek, 2019), a digital repository of verb argument structures manually annotated with Semantic Type labels on their argument slots following a methodology inspired by Corpus Pattern Analysis (Hanks, 2004 & 2013; Hanks & Pustejovsky, 2005). CROATPAS will be made available online in 2020. Semantic Type labelling is not only well-suited to annotate verbal polysemy, but also metonymic shifts in verb argument combinations, which in Generative Lexicon (Pustejovsky, 1995 & 1998; Pustejovsky & Ježek, 2008) are called Semantic Type coercions. From a sub lexical point of view, Semantic Type coercions can be considered as exploitations of one of the qualia roles of those Semantic Types which do not satisfy a verb’s selectional requirements, but do not trigger a different verb sense. Overall, we were able to identify 62 different Semantic Type coercions linked to 1,052 metonymic corpus lines. In the future, we plan to compare our results with those from an equivalent study on Italian verbs (Romani, 2020) for a crosslinguistic analysis of metonymic shifts.
In this paper, we present the ForwardQuestions data set, made of human-generated questions related to knowledge triples. This data set results from the conversion and merger of the existing SimpleDBPediaQA and SimpleQuestionsWikidata data sets, including the mapping of predicates from DBPedia to Wikidata, and the selection of ‘forward’ questions as opposed to ‘backward’ ones. The new data set can be used to generate novel questions given an unseen Wikidata triple, by replacing the subjects of existing questions with the new one and then selecting the best candidate questions using semantic and syntactic criteria. Evaluation results indicate that the question generation method using ForwardQuestions improves the quality of questions by about 20% with respect to a baseline not using ranking criteria.
We present some issues in the development of the semantic annotation of IMAGACT, a multimodal and multilingual ontology of actions. The resource is structured on action concepts that are meant to be cognitive entities and to which a linguistic caption is attached. For each of these concepts, we annotate the minimal thematic structure of the caption and the possible argument alternations allowed. We present some insights on this process with regards to the notion of thematic structure and the relationship between action concepts and linguistic expressions. From the empirical evidence provided by the annotation, we discuss on the very nature of thematic structure, arguing that it is neither a property of the verb itself nor a property of action concepts. We further show what is the relation between thematic structure and 1- the semantic variation of action verbs; 2- the lexical variation of action concepts.
Effective, professional and socially competent dialogue of health care providers with their patients is essential to best practice in medicine. To identify, categorize and quantify salient features of patient-provider communication, to model interactive processes in medical encounters and to design digital interactive medical services, two important instruments have been developed: (1) medical interaction analysis systems with the Roter Interaction Analysis System (RIAS) as the most widely used by medical practitioners and (2) dialogue act annotation schemes with ISO 24617-2 as a multidimensional taxonomy of interoperable semantic concepts widely used for corpus annotation and dialogue systems design. Neither instrument fits all purposes. In this paper, we perform a systematic comparative analysis of the categories defined in the RIAS and ISO taxonomies. Overcoming the deficiencies and gaps that were found, we propose a number of extensions to the ISO annotation scheme, making it a powerful analytical and modelling instrument for the analysis, modelling and assessment of medical communication.
In this paper, we provide the basic guidelines towards the detection and linguistic analysis of events in Kannada. Kannada is a morphologically rich, resource poor Dravidian language spoken in southern India. As most information retrieval and extraction tasks are resource intensive, very little work has been done on Kannada NLP, with almost no efforts in discourse analysis and dataset creation for representing events or other semantic annotations in the text. In this paper, we linguistically analyze what constitutes an event in this language, the challenges faced with discourse level annotation and representation due to the rich derivational morphology of the language that allows free word order, numerous multi-word expressions, adverbial participle constructions and constraints on subject-verb relations. Therefore, this paper is one of the first attempts at a large scale discourse level annotation for Kannada, which can be used for semantic annotation and corpus development for other tasks in the language.
The purpose of this paper is to present a prospective and interdisciplinary research project seeking to ontologize knowledge of the domain of Outsider Art, that is, the art created outside the boundaries of official culture. The goal is to combine ontology engineering methodologies to develop a knowledge base which i) examines the relation between social exclusion and cultural productions, ii) standardizes the terminology of Outsider Art and iii) enables semantic interoperability between cultural metadata relevant to Outsider Art. The Outsider Art ontology will integrate some existing ontologies and terminologies, such as the CIDOC - Conceptual Reference Model (CRM), the Art & Architecture Thesaurus and the Getty Union List of Artist Names, among other resources. Natural Language Processing and Machine Learning techniques will be fundamental instruments for knowledge acquisition and elicitation. NLP techniques will be used to annotate bibliographies of relevant outsider artists and descriptions of outsider artworks with linguistic information. Machine Learning techniques will be leveraged to acquire knowledge from linguistic features embedded in both types of texts.
In this paper we focus on creation of interoperable annotation resources that make up a significant proportion of an on-going project on the development of conceptually annotated multilingual corpora for the domain of terrorist attacks in three languages (English, French and Russian) that can be used for comparative linguistic research, intelligent content and trend analysis, summarization, machine translation, etc. Conceptual annotation is understood as a type of task-oriented domain-specific semantic annotation. The annotation process in our project relies on ontological analysis. The paper details on the issues of the development of both static and dynamic resources such as a universal conceptual annotation scheme, multilingual domain ontology and multipurpose annotation platform with flexible settings, which can be used for the automation of the conceptual resource acquisition and of the annotation process, as well as for the documentation of the annotated corpora specificities. The resources constructed in the course of the research are also to be used for developing concept disambiguation metrics by means of qualitative and quantitative analysis of the golden portion of the conceptually annotated multilingual corpora and of the annotation platform linguistic knowledge.
This paper presents the PORTULAN CLARIN Research Infrastructure for the Science and Technology of Language, which is part of the European research infrastructure CLARIN ERIC as its Portuguese national node, and belongs to the Portuguese National Roadmap of Research Infrastructures of Strategic Relevance. It encompasses a repository, where resources and metadata are deposited for long-term archiving and access, and a workbench, where Language Technology tools and applications are made available through different modes of interaction, among many other services. It is an asset of utmost importance for the technological development of natural languages and for their preparation for the digital age, contributing to ensure the citizenship of their speakers in the information society.
In this paper we describe the current state of development of the Linguistic Linked Open Data (LLOD) infrastructure, an LOD(sub-)cloud of linguistic resources, which covers various linguistic data bases, lexicons, corpora, terminology and metadata repositories.We give in some details an overview of the contributions made by the European H2020 projects “Prêt-à-LLOD” (‘Ready-to-useMultilingual Linked Language Data for Knowledge Services across Sectors’) and “ELEXIS” (‘European Lexicographic Infrastructure’) to the further development of the LLOD.
This paper presents an example architecture for a scalable, secure and resilient Machine Translation (MT) platform, using components available via Amazon Web Services (AWS). It is increasingly common for a single news organisation to publish and monitor news sources in multiple languages. A growth in news sources makes this increasingly challenging and time-consuming but MT can help automate some aspects of this process. Building a translation service provides a single integration point for news room tools that use translation technology allowing MT models to be integrated into a system once, rather than each time the translation technology is needed. By using a range of services provided by AWS, it is possible to architect a platform where multiple pre-existing technologies are combined to build a solution, as opposed to developing software from scratch for deployment on a single virtual machine. This increases the speed at which a platform can be developed and allows the use of well-maintained services. However, a single service also provides challenges. It is key to consider how the platform will scale when handling many users and how to ensure the platform is resilient.
This paper describes the on-going work carried out within the CoBiLiRo (Bimodal Corpus for Romanian Language) research project, part of ReTeRom (Resources and Technologies for Developing Human-Machine Interfaces in Romanian). Data annotation finds increasing use in speech recognition and synthesis with the goal to support learning processes. In this context, a variety of different annotation systems for application to Speech and Text Processing environments have been presented. Even if many designs for the data annotations workflow have emerged, the process of handling metadata, to manage complex user-defined annotations, is not covered enough. We propose a design of the format aimed to serve as an annotation standard for bimodal resources, which facilitates searching, editing and statistical analysis operations over it. The design and implementation of an infrastructure that houses the resources are also presented. The goal is widening the dissemination of bimodal corpora for research valorisation and use in applications. Also, this study reports on the main operations of the web Platform which hosts the corpus and the automatic conversion flows that brings the submitted files at the format accepted by the Platform.
CLARIN is a European Research Infrastructure providing access to digital language resources and tools from across Europe and beyond to researchers in the humanities and social sciences. This paper focuses on CLARIN as a platform for the sharing of language resources. It zooms in on the service offer for the aggregation of language repositories and the value proposition for a number of communities that benefit from the enhanced visibility of their data and services as a result of integration in CLARIN. The enhanced findability of language resources is serving the social sciences and humanities (SSH) community at large and supports research communities that aim to collaborate based on virtual collections for a specific domain. The paper also addresses the wider landscape of service platforms based on language technologies which has the potential of becoming a powerful set of interoperable facilities to a variety of communities of use.
We describe the European Language Resource Infrastructure (ELRI), a decentralised network to help collect, prepare and share language resources. The infrastructure was developed within a project co-funded by the Connecting Europe Facility Programme of the European Union, and has been deployed in the four Member States participating in the project, namely France, Ireland, Portugal and Spain. ELRI provides sustainable and flexible means to collect and share language resources via National Relay Stations, to which members of public institutions can freely subscribe. The infrastructure includes fully automated data processing engines to facilitate the preparation, sharing and wider reuse of useful language resources that can help optimise human and automated translation services in the European Union.
This paper presents our progress towards deploying a versatile communication platform in the task of highly multilingual live speech translation for conferences and remote meetings live subtitling. The platform has been designed with a focus on very low latency and high flexibility while allowing research prototypes of speech and text processing tools to be easily connected, regardless of where they physically run. We outline our architecture solution and also briefly compare it with the ELG platform. Technical details are provided on the most important components and we summarize the test deployment events we ran so far.
Eco is Pangeanic’s customer portal for generic or specialized translation services (machine translation and post-editing, generic API MT and custom API MT). Users can request the processing (translation) of files in different formats. Moreover, a client user can manage the engines and models allowing their cloning and retraining.
We present an software platform and API that combines various ML and NLP approaches for the analysis and enrichment of textual content. The platform’s design and implementation is guided by the goal to allow non-technical users to conduct their own experiments and training runs on their respective data, allowing to test, tune and deploy analysis models for production. Dedicated specific packages for subtasks such as document structure processing, document categorization, annotation with existing thesauri, disambiguation and linking, annotation with newly created entity recognizers and summarization – available as open source components in isolation – are combined into an end-user-facing, collaborative, scalable platform to support large-scale industrial document analysis document analysis. We see the Sherpa’s setup as an answer to the observation that ML has reached a level of maturity that allows to attain useful results in many analysis scenarios today, but that in-depth technical competencies in the required fields of NLP and AI is often scarce; a setup that focusses on non-technical domain-expert end-users can help to bring required analysis functionalities closer to the day-to-day reality in business contexts.
Several web services for various natural language processing (NLP) tasks (‘‘NLP-as-a-service” or NLPaaS) have recently been made publicly available. However, despite their similar functionality these services often differ in the protocols they use, thus complicating the development of clients accessing them. A survey of currently available NLPaaS services suggests that it may be possible to identify a minimal application layer protocol that can be shared by NLPaaS services without sacrificing functionality or convenience, while at the same time simplifying the development of clients for these services. In this paper, we hope to raise awareness of the interoperability problems caused by the variety of existing web service protocols, and describe an effort to identify a set of best practices for NLPaaS protocol design. To that end, we survey and compare protocols used by NLPaaS services and suggest how these protocols may be further aligned to reduce variation.
Nowadays the scarcity and dispersion of open-source NLP resources and tools in and for African languages make it difficult for researchers to truly fit these languages into current algorithms of artificial intelligence, resulting in the stagnation of these numerous languages, as far as technological progress is concerned. Created in 2017, with the aim of building communities of voluntary contributors around African native and/or national languages, cultures, NLP technologies and artificial intelligence, the NTeALan association has set up a series of web collaborative platforms intended to allow the aforementioned communities to create and manage their own lexicographic and linguistic resources. This paper aims at presenting the first versions of three lexicographic platforms that we developed in and for African languages: the REST/GraphQL API for saving lexicographic resources, the dictionary management platform and the collaborative dictionary platform. We also describe the data representation format used for these resources. After experimenting with a few dictionaries and looking at users feedback, we are convinced that only collaboration-based approaches and platforms can effectively respond to challenges of producing quality resources in and for African native and/or national languages.
We present a workflow manager for the flexible creation and customisation of NLP processing pipelines. The workflow manager addresses challenges in interoperability across various different NLP tasks and hardware-based resource usage. Based on the four key principles of generality, flexibility, scalability and efficiency, we present the first version of the workflow manager by providing details on its custom definition language, explaining the communication components and the general system architecture and setup. We currently implement the system, which is grounded and motivated by real-world industry use cases in several innovation and transfer projects.
This paper presents RELATE (http://relate.racai.ro), a high-performance natural language platform designed for Romanian language. It is meant both for demonstration of available services, from text-span annotations to syntactic dependency trees as well as playing or automatically synthesizing Romanian words, and for the development of new annotated corpora. It also incorporates the search engines for the large COROLA reference corpus of contemporary Romanian and the Romanian wordnet. It integrates multiple text and speech processing modules and exposes their functionality through a web interface designed for the linguist researcher. It makes use of a scheduler-runner architecture, allowing processing to be distributed across multiple computing nodes. A series of input/output converters allows large corpora to be loaded, processed and exported according to user preferences.
In this paper, we present LinTO, an intelligent voice platform and smart room assistant for improving efficiency and productivity in business. Our objective is to build a Spoken Language Understanding system that maintains high performance in both Automatic Speech Recognition (ASR) and Natural Language Processing while being portable and scalable. In this paper we describe the LinTO architecture and our approach to ASR engine training which takes advantage of recent advances in deep learning while guaranteeing high-performance real-time processing. Unlike the existing solutions, the LinTO platform is open source for commercial and non-commercial use
With regard to the wider area of AI/LT platform interoperability, we concentrate on two core aspects: (1) cross-platform search and discovery of resources and services; (2) composition of cross-platform service workflows. We devise five different levels (of increasing complexity) of platform interoperability that we suggest to implement in a wider federation of AI/LT platforms. We illustrate the approach using the five emerging AI/LT platforms AI4EU, ELG, Lynx, QURATOR and SPEAKER.
This paper presents the COMPRISE cloud platform that is developed in the H2020 project. We present an overview of the COMPRISE project, its main goals, components, and how the cloud platform fits in the context of the overall project. The COMPRISE cloud platform is presented in more detail – main users, use scenarios, functions, implementation details, and how it will be used by both COMPRISE’s targeted audience and the broader language-technology community.
This paper describes the workflow and architecture adopted by a linguistic research project. We report our experience and present the research outputs turned into resources that we wish to share with the community. We discuss the current limitations and the next steps that could be taken for the scaling and development of our research project. Allying NLP and language-centric AI, we discuss similar projects and possible ways to start collaborating towards potential platform interoperability.
To process the syntactic structures of a language in ways that are compatible with human expectations, we need computational representations of lexical and syntactic properties that form the basis of human knowledge of words and sentences. Recent neural-network-based and distributed semantics techniques have developed systems of considerable practical success and impressive performance. As has been advocated by many, however, such systems still lack human-like properties. In particular, linguistic, psycholinguistic and neuroscientific investigations have shown that human processing of sentences is sensitive to structure and unbounded relations. In the spirit of better understanding the structure building and long-distance properties of neural networks, I will present an overview of recent results on agreement and island effects in syntax in several languages. While certain sets of results in the literature indicate that neural language models exhibit long-distance agreement abilities, other finer-grained investigation of how these effects are calculated indicates that that the similarity spaces they define do not correlate with human experimental results on intervention similarity in long-distance dependencies. This opens the way to reflections on how to better match the syntactic properties of natural languages in the representations of neural models.
The carbon footprint of natural language processing research has been increasing in recent years due to its reliance on large and inefficient neural network implementations. Distillation is a network compression technique which attempts to impart knowledge from a large model to a smaller one. We use teacher-student distillation to improve the efficiency of the Biaffine dependency parser which obtains state-of-the-art performance with respect to accuracy and parsing speed (Dozat and Manning, 2017). When distilling to 20% of the original model’s trainable parameters, we only observe an average decrease of ∼1 point for both UAS and LAS across a number of diverse Universal Dependency treebanks while being 2.30x (1.19x) faster than the baseline model on CPU (GPU) at inference time. We also observe a small increase in performance when compressing to 80% for some treebanks. Finally, through distillation we attain a parser which is not only faster but also more accurate than the fastest modern parser on the Penn Treebank.
We present a neural end-to-end architecture for negation resolution based on a formulation of the task as a graph parsing problem. Our approach allows for the straightforward inclusion of many types of graph-structured features without the need for representation-specific heuristics. In our experiments, we specifically gauge the usefulness of syntactic information for negation resolution. Despite the conceptual simplicity of our architecture, we achieve state-of-the-art results on the Conan Doyle benchmark dataset, including a new top result for our best model.
Graph-based and transition-based dependency parsers used to have different strengths and weaknesses. Therefore, combining the outputs of parsers from both paradigms used to be the standard approach to improve or analyze their performance. However, with the recent adoption of deep contextualized word representations, the chief weakness of graph-based models, i.e., their limited scope of features, has been mitigated. Through two popular combination techniques – blending and stacking – we demonstrate that the remaining diversity in the parsing models is reduced below the level of models trained with different random seeds. Thus, an integration no longer leads to increased accuracy. When both parsers depend on BiLSTMs, the graph-based architecture has a consistent advantage. This advantage stems from globally-trained BiLSTM representations, which capture more distant look-ahead syntactic relations. Such representations can be exploited through multi-task learning, which improves the transition-based parser, especially on treebanks with a high ratio of right-headed dependencies.
We propose an end-to-end variational autoencoding parsing (VAP) model for semi-supervised graph-based projective dependency parsing. It encodes the input using continuous latent variables in a sequential manner by deep neural networks (DNN) that can utilize the contextual information, and reconstruct the input using a generative model. The VAP model admits a unified structure with different loss functions for labeled and unlabeled data with shared parameters. We conducted experiments on the WSJ data sets, showing the proposed model can use the unlabeled data to increase the performance on a limited amount of labeled data, on a par with a recently proposed semi-supervised parser with faster inference.
Syntactic surprisal has been shown to have an effect on human sentence processing, and can be predicted from prefix probabilities of generative incremental parsers. Recent state-of-the-art incremental generative neural parsers are able to produce accurate parses and surprisal values but have unbounded stack memory, which may be used by the neural parser to maintain explicit in-order representations of all previously parsed words, inconsistent with results of human memory experiments. In contrast, humans seem to have a bounded working memory, demonstrated by inhibited performance on word recall in multi-clause sentences (Bransford and Franks, 1971), and on center-embedded sentences (Miller and Isard,1964). Bounded statistical parsers exist, but are less accurate than neural parsers in predict-ing reading times. This paper describes a neural incremental generative parser that is able to provide accurate surprisal estimates and can be constrained to use a bounded stack. Results show that the accuracy gains of neural parsers can be reliably extended to psycholinguistic modeling without risk of distortion due to un-bounded working memory.
The goal of homomorphic encryption is to encrypt data such that another party can operate on it without being explicitly exposed to the content of the original data. We introduce an idea for a privacy-preserving transformation on natural language data, inspired by homomorphic encryption. Our primary tool is obfuscation, relying on the properties of natural language. Specifically, a given English text is obfuscated using a neural model that aims to preserve the syntactic relationships of the original sentence so that the obfuscated sentence can be parsed instead of the original one. The model works at the word level, and learns to obfuscate each word separately by changing it into a new word that has a similar syntactic role. The text obfuscated by our model leads to better performance on three syntactic parsers (two dependency and one constituency parsers) in comparison to an upper-bound random substitution baseline. More specifically, the results demonstrate that as more terms are obfuscated (by their part of speech), the substitution upper bound significantly degrades, while the neural model maintains a relatively high performing parser. All of this is done without much sacrifice of privacy compared to the random substitution upper bound. We also further analyze the results, and discover that the substituted words have similar syntactic properties, but different semantic content, compared to the original words.
Semiring parsing is an elegant framework for describing parsers by using semiring weighted logic programs. In this paper we present a generalization of this concept: latent-variable semiring parsing. With our framework, any semiring weighted logic program can be latentified by transforming weights from scalar values of a semiring to rank-n arrays, or tensors, of semiring values, allowing the modelling of latent-variable models within the semiring parsing framework. Semiring is too strong a notion when dealing with tensors, and we have to resort to a weaker structure: a partial semiring. We prove that this generalization preserves all the desired properties of the original semiring framework while strictly increasing its expressiveness.
We present new experiments that transfer techniques from Probabilistic Context-free Grammars with Latent Annotations (PCFG-LA) to two grammar formalisms for discontinuous parsing: linear context-free rewriting systems and hybrid grammars. In particular, Dirichlet priors during EM training, ensemble models, and a new nonterminal scheme for hybrid grammars are evaluated. We find that our grammars are more accurate than previous approaches based on discontinuous grammar formalisms and early instances of the discriminative models but inferior to recent discriminative parsers.
In the field of constituent parsing, probabilistic grammar formalisms have been studied to model the syntactic structure of natural language. More recently, approaches utilizing neural models gained lots of traction in this field, as they achieved accurate results at high speed. We aim for a symbiosis between probabilistic linear context-free rewriting systems (PLCFRS) as a probabilistic grammar formalism and neural models to get the best of both worlds: the interpretability of grammars, and the speed and accuracy of neural models. To combine these two, we consider the approach of supertagging that requires lexicalized grammar formalisms. Here, we present a procedure which turns any PLCFRS G into an equivalent lexicalized PLCFRS G’. The derivation trees in G’ are then mapped to equivalent derivations in G. Our construction for G’ preserves the probability assignment and does not increase parsing complexity compared to G.
Neural unsupervised parsing (UP) models learn to parse without access to syntactic annotations, while being optimized for another task like language modeling. In this work, we propose self-training for neural UP models: we leverage aggregated annotations predicted by copies of our model as supervision for future copies. To be able to use our model’s predictions during training, we extend a recent neural UP architecture, the PRPN (Shen et al., 2018a), such that it can be trained in a semi-supervised fashion. We then add examples with parses predicted by our model to our unlabeled UP training data. Our self-trained model outperforms the PRPN by 8.1% F1 and the previous state of the art by 1.6% F1. In addition, we show that our architecture can also be helpful for semi-supervised parsing in ultra-low-resource settings.
The earliest models for discontinuous constituency parsers used mildly context-sensitive grammars, but the fashion has changed in recent years to grammar-less transition-based parsers that use strong neural probabilistic models to greedily predict transitions. We argue that grammar-based approaches still have something to contribute on top of what is offered by transition-based parsers. Concretely, by using a grammar formalism to restrict the space of possible trees we can use dynamic programming parsing algorithms for exact search for the most probable tree. Previous chart-based parsers for discontinuous formalisms used probabilistically weak generative models. We instead use a span-based discriminative neural model that preserves the dynamic programming properties of the chart parsers. Our parser does not use an explicit grammar, but it does use explicit grammar formalism constraints: we generate only trees that are within the LCFRS-2 formalism. These properties allow us to construct a new parsing algorithm that runs in lower worst-case time complexity of O(l nˆ4 +nˆ6), where n is the sentence length and l is the number of unique non-terminal labels. This parser is efficient in practice, provides best results among chart-based parsers, and is competitive with the best transition based parsers. We also show that the main bottleneck for further improvement in performance is in the restriction of fan-out to degree 2. We show that well-nestedness is helpful in speeding up parsing, but lowers accuracy.
In this paper, we first open on important issues regarding the Penn Korean Universal Treebank (PKT-UD) and address these issues by revising the entire corpus manually with the aim of producing cleaner UD annotations that are more faithful to Korean grammar. For compatibility to the rest of UD corpora, we follow the UDv2 guidelines, and extensively revise the part-of-speech tags and the dependency relations to reflect morphological features and flexible word- order aspects in Korean. The original and the revised versions of PKT-UD are experimented with transformer-based parsing models using biaffine attention. The parsing model trained on the revised corpus shows a significant improvement of 3.0% in labeled attachment score over the model trained on the previous corpus. Our error analysis demonstrates that this revision allows the parsing model to learn relations more robustly, reducing several critical errors that used to be made by the previous model.
This paper presents the development of a deep parser for Spanish that uses a HPSG grammar and returns trees that contain both syntactic and semantic information. The parsing process uses a top-down approach implemented using LSTM neural networks, and achieves good performance results in terms of syntactic constituency and dependency metrics, and also SRL. We describe the grammar, corpus and implementation of the parser. Our process outperforms a CKY baseline and other Spanish parsers in terms of global metrics and also for some specific Spanish phenomena, such as clitics reduplication and relative referents.
Recent progress in grammar induction has shown that grammar induction is possible without explicit assumptions of language specific knowledge. However, evaluation of induced grammars usually has ignored phrasal labels, an essential part of a grammar. Experiments in this work using a labeled evaluation metric, RH, show that linguistically motivated predictions about grammar sparsity and use of categories can only be revealed through labeled evaluation. Furthermore, depth-bounding as an implementation of human memory constraints in grammar inducers is still effective with labeled evaluation on multilingual transcribed child-directed utterances.
This overview introduces the task of parsing into enhanced universal dependencies, describes the datasets used for training and evaluation, and evaluation metrics. We outline various approaches and discuss the results of the shared task.
We present the approach of the TurkuNLP group to the IWPT 2020 shared task on Multilingual Parsing into Enhanced Universal Dependencies. The task involves 28 treebanks in 17 different languages and requires parsers to generate graph structures extending on the basic dependency trees. Our approach combines language-specific BERT models, the UDify parser, neural sequence-to-sequence lemmatization and a graph transformation approach encoding the enhanced structure into a dependency tree. Our submission averaged 84.5% ELAS, ranking first in the shared task. We make all methods and resources developed for this study freely available under open licenses from https://turkunlp.org.
This paper describes our system to predict enhanced dependencies for Universal Dependencies (UD) treebanks, which ranked 2nd in the Shared Task on Enhanced Dependency Parsing with an average ELAS of 82.60%. Our system uses a hybrid two-step approach. First, we use a graph-based parser to extract a basic syntactic dependency tree. Then, we use a set of linguistic rules which generate the enhanced dependencies for the syntactic tree. The application of these rules is optimized using a classifier which predicts their suitability in the given context. A key advantage of this approach is its language independence, as rules rely solely on dependency trees and UPOS tags which are shared across all languages.
This paper presents our enhanced dependency parsing approach using transformer encoders, coupled with a simple yet powerful ensemble algorithm that takes advantage of both tree and graph dependency parsing. Two types of transformer encoders are compared, a multilingual encoder and language-specific encoders. Our dependency tree parsing (DTP) approach generates only primary dependencies to form trees whereas our dependency graph parsing (DGP) approach handles both primary and secondary dependencies to form graphs. Since DGP does not guarantee the generated graphs are acyclic, the ensemble algorithm is designed to add secondary arcs predicted by DGP to primary arcs predicted by DTP. Our results show that models using the multilingual encoder outperform ones using the language specific encoders for most languages. The ensemble models generally show higher labeled attachment score on enhanced dependencies (ELAS) than the DTP and DGP models. As the result, our best models rank the third place on the macro-average ELAS over 17 languages.
We present the system submission from the FASTPARSE team for the EUD Shared Task at IWPT 2020. We engaged with the task by focusing on efficiency. For this we considered training costs and inference efficiency. Our models are a combination of distilled neural dependency parsers and a rule-based system that projects UD trees into EUD graphs. We obtained an average ELAS of 74.04 for our official submission, ranking 4th overall.
To accomplish the shared task on dependency parsing we explore the use of a linear transition-based neural dependency parser as well as a combination of three of them by means of a linear tree combination algorithm. We train separate models for each language on the shared task data. We compare our base parser with two biaffine parsers and also present an ensemble combination of all five parsers, which achieves an average UAS 1.88 point lower than the top official submission. For producing the enhanced dependencies, we exploit a hybrid approach, coupling an algorithmic graph transformation of the dependency tree with predictions made by a multitask machine learning model.
This paper presents the system used in our submission to the IWPT 2020 Shared Task. Our system is a graph-based parser with second-order inference. For the low-resource Tamil corpora, we specially mixed the training data of Tamil with other languages and significantly improved the performance of Tamil. Due to our misunderstanding of the submission requirements, we submitted graphs that are not connected, which makes our system only rank 6th over 10 teams. However, after we fixed this problem, our system is 0.6 ELAS higher than the team that ranked 1st in the official results.
In this paper, we present the submission of team CLASP to the IWPT 2020 Shared Task on parsing enhanced universal dependencies. We develop a tree-to-graph transformation algorithm based on dependency patterns. This algorithm can transform gold UD trees to EUD graphs with an ELAS score of 81.55 and a EULAS score of 96.70. These results show that much of the information needed to construct EUD graphs from UD trees are present in the UD trees. Coupled with a standard UD parser, the method applies to the official test data and yields and ELAS score of 67.85 and a EULAS score is 80.18.
We describe the ADAPT system for the 2020 IWPT Shared Task on parsing enhanced Universal Dependencies in 17 languages. We implement a pipeline approach using UDPipe and UDPipe-future to provide initial levels of annotation. The enhanced dependency graph is either produced by a graph-based semantic dependency parser or is built from the basic tree using a small set of heuristics. Our results show that, for the majority of languages, a semantic dependency parser can be successfully applied to the task of parsing enhanced dependencies. Unfortunately, we did not ensure a connected graph as part of our pipeline approach and our competition submission relied on a last-minute fix to pass the validation script which harmed our official evaluation scores significantly. Our submission ranked eighth in the official evaluation with a macro-averaged coarse ELAS F1 of 67.23 and a treebank average of 67.49. We later implemented our own graph-connecting fix which resulted in a score of 79.53 (language average) or 79.76 (treebank average), which would have placed fourth in the competition evaluation.
We present Køpsala, the Copenhagen-Uppsala system for the Enhanced Universal Dependencies Shared Task at IWPT 2020. Our system is a pipeline consisting of off-the-shelf models for everything but enhanced graph parsing, and for the latter, a transition-based graph parser adapted from Che et al. (2019). We train a single enhanced parser model per language, using gold sentence splitting and tokenization for training, and rely only on tokenized surface forms and multilingual BERT for encoding. While a bug introduced just before submission resulted in a severe drop in precision, its post-submission fix would bring us to 4th place in the official ranking, according to average ELAS. Our parser demonstrates that a unified pipeline is effective for both Meaning Representation Parsing and Enhanced Universal Dependencies.
This paper presents our system at the IWPT 2020 Shared Task on Parsing into Enhanced Universal Dependencies. Using a biaffine classifier architecture (Dozat and Manning, 2017) which operates directly on finetuned RoBERTa embeddings, our parser generates enhanced UD graphs by predicting the best dependency label (or absence of a dependency) for each pair of tokens in the sentence. We address label sparsity issues by replacing lexical items in relations with placeholders at prediction time, later retrieving them from the parse in a rule-based fashion. In addition, we ensure structural graph constraints using a simple set of heuristics. On the English blind test data, our system achieves a very high parsing accuracy, ranking 1st out of 10 with an ELAS F1 score of 88.94%.
The evaluation campaign of the International Conference on Spoken Language Translation (IWSLT 2020) featured this year six challenge tracks: (i) Simultaneous speech translation, (ii) Video speech translation, (iii) Offline speech translation, (iv) Conversational speech translation, (v) Open domain translation, and (vi) Non-native speech translation. A total of teams participated in at least one of the tracks. This paper introduces each track’s goal, data and evaluation metrics, and reports the results of the received submissions.
This paper describes the ON-TRAC Consortium translation systems developed for two challenge tracks featured in the Evaluation Campaign of IWSLT 2020, offline speech translation and simultaneous speech translation. ON-TRAC Consortium is composed of researchers from three French academic laboratories: LIA (Avignon Université), LIG (Université Grenoble Alpes), and LIUM (Le Mans Université). Attention-based encoder-decoder models, trained end-to-end, were used for our submissions to the offline speech translation track. Our contributions focused on data augmentation and ensembling of multiple models. In the simultaneous speech translation track, we build on Transformer-based wait-k models for the text-to-text subtask. For speech-to-text simultaneous translation, we attach a wait-k MT system to a hybrid ASR system. We propose an algorithm to control the latency of the ASR+MT cascade and achieve a good latency-quality trade-off on both subtasks.
AppTek and RWTH Aachen University team together to participate in the offline and simultaneous speech translation tracks of IWSLT 2020. For the offline task, we create both cascaded and end-to-end speech translation systems, paying attention to careful data selection and weighting. In the cascaded approach, we combine high-quality hybrid automatic speech recognition (ASR) with the Transformer-based neural machine translation (NMT). Our end-to-end direct speech translation systems benefit from pretraining of adapted encoder and decoder components, as well as synthetic data and fine-tuning and thus are able to compete with cascaded systems in terms of MT quality. For simultaneous translation, we utilize a novel architecture that makes dynamic decisions, learned from parallel data, to determine when to continue feeding on input or generate output words. Experiments with speech and text input show that even at low latency this architecture leads to superior translation results.
This paper describes KIT’s submissions to the IWSLT2020 Speech Translation evaluation campaign. We first participate in the simultaneous translation task, in which our simultaneous models are Transformer based and can be efficiently trained to obtain low latency with minimized compromise in quality. On the offline speech translation task, we applied our new Speech Transformer architecture to end-to-end speech translation. The obtained model can provide translation quality which is competitive to a complicated cascade. The latter still has the upper hand, thanks to the ability to transparently access to the transcription, and resegment the inputs to avoid fragmentation.
In this paper, we describe end-to-end simultaneous speech-to-text and text-to-text translation systems submitted to IWSLT2020 online translation challenge. The systems are built by adding wait-k and meta-learning approaches to the Transformer architecture. The systems are evaluated on different latency regimes. The simultaneous text-to-text translation achieved a BLEU score of 26.38 compared to the competition baseline score of 14.17 on the low latency regime (Average latency ≤ 3). The simultaneous speech-to-text system improves the BLEU score by 7.7 points over the competition baseline for the low latency regime (Average Latency ≤ 1000).
This paper describes the system that was submitted by DiDi Labs to the offline speech translation task for IWSLT 2020. We trained an end-to-end system that translates audio from English TED talks to German text, without producing intermediate English text. We use the S-Transformer architecture and train using the MuSTC dataset. We also describe several additional experiments that were attempted, but did not yield improved results.
In this paper, we describe the system submitted to the IWSLT 2020 Offline Speech Translation Task. We adopt the Transformer architecture coupled with the meta-learning approach to build our end-to-end Speech-to-Text Translation (ST) system. Our meta-learning approach tackles the data scarcity of the ST task by leveraging the data available from Automatic Speech Recognition (ASR) and Machine Translation (MT) tasks. The meta-learning approach combined with synthetic data augmentation techniques improves the model performance significantly and achieves BLEU scores of 24.58, 27.51, and 27.61 on IWSLT test 2015, MuST-C test, and Europarl-ST test sets respectively.
This paper describes FBK’s participation in the IWSLT 2020 offline speech translation (ST) task. The task evaluates systems’ ability to translate English TED talks audio into German texts. The test talks are provided in two versions: one contains the data already segmented with automatic tools and the other is the raw data without any segmentation. Participants can decide whether to work on custom segmentation or not. We used the provided segmentation. Our system is an end-to-end model based on an adaptation of the Transformer for speech data. Its training process is the main focus of this paper and it is based on: i) transfer learning (ASR pretraining and knowledge distillation), ii) data augmentation (SpecAugment, time stretch and synthetic data), iii)combining synthetic and real data marked as different domains, and iv) multi-task learning using the CTC loss. Finally, after the training with word-level knowledge distillation is complete, our ST models are fine-tuned using label smoothed cross entropy. Our best model scored 29 BLEU on the MuST-CEn-De test set, which is an excellent result compared to recent papers, and 23.7 BLEU on the same data segmented with VAD, showing the need for researching solutions addressing this specific data condition.
We took part in the offline End-to-End English to German TED lectures translation task. We based our solution on our last year’s submission. We used a slightly altered Transformer architecture with ResNet-like convolutional layer preparing the audio input to Transformer encoder. To improve the model’s quality of translation we introduced two regularization techniques and trained on machine translated Librispeech corpus in addition to iwslt-corpus, TEDLIUM2 andMust_C corpora. Our best model scored almost 3 BLEU higher than last year’s model. To segment 2020 test set we used exactly the same procedure as last year.
This paper describes the University of Helsinki Language Technology group’s participation in the IWSLT 2020 offline speech translation task, addressing the translation of English audio into German text. In line with this year’s task objective, we train both cascade and end-to-end systems for spoken language translation. We opt for an end-to-end multitasking architecture with shared internal representations and a cascade approach that follows a standard procedure consisting of ASR, correction, and MT stages. We also describe the experiments that served as a basis for the submitted systems. Our experiments reveal that multitasking training with shared internal representations is not only possible but allows for knowledge-transfer across modalities.
This report summarizes the Air Force Research Laboratory (AFRL) submission to the offline spoken language translation (SLT) task as part of the IWSLT 2020 evaluation campaign. As in previous years, we chose to adopt the cascade approach of using separate systems to perform speech activity detection, automatic speech recognition, sentence segmentation, and machine translation. All systems were neural based, including a fully-connected neural network for speech activity detection, a Kaldi factorized time delay neural network with recurrent neural network (RNN) language model rescoring for speech recognition, a bidirectional RNN with attention mechanism for sentence segmentation, and transformer networks trained with OpenNMT and Marian for machine translation. Our primary submission yielded BLEU scores of 21.28 on tst2019 and 23.33 on tst2020.
This paper describes the LIT Team’s submission to the IWSLT2020 open domain translation task, focusing primarily on Japanese-to-Chinese translation direction. Our system is based on the organizers’ baseline system, but we do more works on improving the Transform baseline system by elaborate data pre-processing. We manage to obtain significant improvements, and this paper aims to share some data processing experiences in this translation task. Large-scale back-translation on monolingual corpus is also investigated. In addition, we also try shared and exclusive word embeddings, compare different granularity of tokens like sub-word level. Our Japanese-to-Chinese translation system achieves a performance of BLEU=34.0 and ranks 2nd among all participating systems.
In this paper, we demonstrate our machine translation system applied for the Chinese-Japanese bidirectional translation task (aka. open domain translation task) for the IWSLT 2020. Our model is based on Transformer (Vaswani et al., 2017), with the help of many popular, widely proved effective data preprocessing and augmentation methods. Experiments show that these methods can improve the baseline model steadily and significantly.
This paper describes the University of Edinburgh’s neural machine translation systems submitted to the IWSLT 2020 open domain Japanese↔Chinese translation task. On top of commonplace techniques like tokenisation and corpus cleaning, we explore character mapping and unsupervised decoding-time adaptation. Our techniques focus on leveraging the provided data, and we show the positive impact of each technique through the gradual improvement of BLEU.
This paper describes the CASIA’s system for the IWSLT 2020 open domain translation task. This year we participate in both Chinese→Japanese and Japanese→Chinese translation tasks. Our system is neural machine translation system based on Transformer model. We augment the training data with knowledge distillation and back translation to improve the translation performance. Domain data classification and weighted domain model ensemble are introduced to generate the final translation result. We compare and analyze the performance on development data with different model settings and different data processing techniques.
We present in this report our submission to IWSLT 2020 Open Domain Translation Task. We built a data pre-processing pipeline to efficiently handle large noisy web-crawled corpora, which boosts the BLEU score of a widely used transformer model in this translation task. To tackle the open-domain nature of this task, back- translation is applied to further improve the translation performance.
In this paper, we introduce University of Tsukuba’s submission to the IWSLT20 Open Domain Translation Task. We participate in both Chinese→Japanese and Japanese→Chinese directions. For both directions, our machine translation systems are based on the Transformer architecture. Several techniques are integrated in order to boost the performance of our models: data filtering, large-scale noised training, model ensemble, reranking and postprocessing. Consequently, our efforts achieve 33.0 BLEU scores for Chinese→Japanese translation and 32.3 BLEU scores for Japanese→Chinese translation.
This paper describes the Xiaomi’s submissions to the IWSLT20 shared open domain translation task for Chinese<->Japanese language pair. We explore different model ensembling strategies based on recent Transformer variants. We also further strengthen our systems via some effective techniques, such as data filtering, data selection, tagged back translation, domain adaptation, knowledge distillation, and re-ranking. Our resulting Chinese->Japanese primary system ranked second in terms of character-level BLEU score among all submissions. Our resulting Japanese->Chinese primary system also achieved a competitive performance.
This paper introduces technical details of machine translation system of Institute of Scientific and Technical Information of China (ISTIC) for the 17th International Conference on Spoken Language Translation (IWSLT 2020). ISTIC participated in both translation tasks of the Open Domain Translation track: Japanese-to-Chinese MT task and Chinese-to-Japanese MT task. The paper mainly elaborates on the model framework, data preprocessing methods and decoding strategies adopted in our system. In addition, the system performance on the development set are given under different settings.
This paper describes Octanove Labs’ submission to the IWSLT 2020 open domain translation challenge. In order to build a high-quality Japanese-Chinese neural machine translation (NMT) system, we use a combination of 1) parallel corpus filtering and 2) back-translation. We have shown that, by using heuristic rules and learned classifiers, the size of the parallel data can be reduced by 70% to 90% without much impact on the final MT performance. We have also shown that including the artificially generated parallel data through back-translation further boosts the metric by 17% to 27%, while self-training contributes little. Aside from a small number of parallel sentences annotated for filtering, no external resources have been used to build our system.
This paper describes NAIST’s NMT system submitted to the IWSLT 2020 conversational speech translation task. We focus on the translation disfluent speech transcripts that include ASR errors and non-grammatical utterances. We tried a domain adaptation method by transferring the styles of out-of-domain data (United Nations Parallel Corpus) to be like in-domain data (Fisher transcripts). Our system results showed that the NMT model with domain adaptation outperformed a baseline. In addition, slight improvement by the style transfer was observed.
Machine translation systems perform reasonably well when the input is well-formed speech or text. Conversational speech is spontaneous and inherently consists of many disfluencies. Producing fluent translations of disfluent source text would typically require parallel disfluent to fluent training data. However, fluent translations of spontaneous speech are an additional resource that is tedious to obtain. This work describes the submission of IIT Bombay to the Conversational Speech Translation challenge at IWSLT 2020. We specifically tackle the problem of disfluency removal in disfluent-to-fluent text-to-text translation assuming no access to fluent references during training. Common patterns of disfluency are extracted from disfluent references and a noise induction model is used to simulate them starting from a clean monolingual corpus. This synthetically constructed dataset is then considered as a proxy for labeled data during training. We also make use of additional fluent text in the target language to help generate fluent translations. This work uses no fluent references during training and beats a baseline model by a margin of 4.21 and 3.11 BLEU points where the baseline uses disfluent and fluent references, respectively. Index Terms- disfluency removal, machine translation, noise induction, leveraging monolingual data, denoising for disfluency removal.
The paper presents details of our system in the IWSLT Video Speech Translation evaluation. The system works in a cascade form, which contains three modules: 1) A proprietary ASR system. 2) A disfluency correction system aims to remove interregnums or other disfluent expressions with a fine-tuned BERT and a series of rule-based algorithms. 3) An NMT System based on the Transformer and trained with massive publicly available corpus.
In this paper, we present our submission to the Non-Native Speech Translation Task for IWSLT 2020. Our main contribution is a proposed speech recognition pipeline that consists of an acoustic model and a phoneme-to-grapheme model. As an intermediate representation, we utilize phonemes. We demonstrate that the proposed pipeline surpasses commercially used automatic speech recognition (ASR) and submit it into the ASR track. We complement this ASR with off-the-shelf MT systems to take part also in the speech translation track.
This paper is an ELITR system submission for the non-native speech translation task at IWSLT 2020. We describe systems for offline ASR, real-time ASR, and our cascaded approach to offline SLT and real-time SLT. We select our primary candidates from a pool of pre-existing systems, develop a new end-to-end general ASR system, and a hybrid ASR trained on non-native speech. The provided small validation set prevents us from carrying out a complex validation, but we submit all the unselected candidates for contrastive evaluation on the test set.
Subtitling is becoming increasingly important for disseminating information, given the enormous amounts of audiovisual content becoming available daily. Although Neural Machine Translation (NMT) can speed up the process of translating audiovisual content, large manual effort is still required for transcribing the source language, and for spotting and segmenting the text into proper subtitles. Creating proper subtitles in terms of timing and segmentation highly depends on information present in the audio (utterance duration, natural pauses). In this work, we explore two methods for applying Speech Translation (ST) to subtitling, a) a direct end-to-end and b) a classical cascade approach. We discuss the benefit of having access to the source language speech for improving the conformity of the generated subtitles to the spatial and temporal subtitling constraints and show that length is not the answer to everything in the case of subtitling-oriented ST.
There has been great progress in improving streaming machine translation, a simultaneous paradigm where the system appends to a growing hypothesis as more source content becomes available. We study a related problem in which revisions to the hypothesis beyond strictly appending words are permitted. This is suitable for applications such as live captioning an audio feed. In this setting, we compare custom streaming approaches to re-translation, a straightforward strategy where each new source token triggers a distinct translation from scratch. We find re-translation to be as good or better than state-of-the-art streaming systems, even when operating under constraints that allow very few revisions. We attribute much of this success to a previously proposed data-augmentation technique that adds prefix-pairs to the training data, which alongside wait-k inference forms a strong baseline for streaming translation. We also highlight re-translation’s ability to wrap arbitrarily powerful MT systems with an experiment showing large improvements from an upgrade to its base model.
Simultaneous machine translation systems rely on a policy to schedule read and write operations in order to begin translating a source sentence before it is complete. In this paper, we demonstrate the use of Adaptive Computation Time (ACT) as an adaptive, learned policy for simultaneous machine translation using the transformer model and as a more numerically stable alternative to Monotonic Infinite Lookback Attention (MILk). We achieve state-of-the-art results in terms of latency-quality tradeoffs. We also propose a method to use our model on unsegmented input, i.e. without sentence boundaries, simulating the condition of translating output from automatic speech recognition. We present first benchmark results on this task.
In simultaneous machine translation, the objective is to determine when to produce a partial translation given a continuous stream of source words, with a trade-off between latency and quality. We propose a neural machine translation (NMT) model that makes dynamic decisions when to continue feeding on input or generate output words. The model is composed of two main components: one to dynamically decide on ending a source chunk, and another that translates the consumed chunk. We train the components jointly and in a manner consistent with the inference conditions. To generate chunked training data, we propose a method that utilizes word alignment while also preserving enough context. We compare models with bidirectional and unidirectional encoders of different depths, both on real speech and text input. Our results on the IWSLT 2020 English-to-German task outperform a wait-k baseline by 2.6 to 3.7% BLEU absolute.
Automatic speech recognition (ASR) systems are primarily evaluated on transcription accuracy. However, in some use cases such as subtitling, verbatim transcription would reduce output readability given limited screen size and reading time. Therefore, this work focuses on ASR with output compression, a task challenging for supervised approaches due to the scarcity of training data. We first investigate a cascaded system, where an unsupervised compression model is used to post-edit the transcribed speech. We then compare several methods of end-to-end speech recognition under output length constraints. The experiments show that with limited data far less than needed for training a model from scratch, we can adapt a Transformer-based ASR model to incorporate both transcription and compression capabilities. Furthermore, the best performance in terms of WER and ROUGE scores is achieved by explicitly modeling the length constraints within the end-to-end ASR system.
We present enhancements to a speech-to-speech translation pipeline in order to perform automatic dubbing. Our architecture features neural machine translation generating output of preferred length, prosodic alignment of the translation with the original speech segments, neural text-to-speech with fine tuning of the duration of each utterance, and, finally, audio rendering to enriches text-to-speech output with background noise and reverberation extracted from the original audio. We report and discuss results of a first subjective evaluation of automatic dubbing of excerpts of TED Talks from English into Italian, which measures the perceived naturalness of automatic dubbing and the relative importance of each proposed enhancement.
A variety of natural language tasks require processing of textual data which contains a mix of natural language and formal languages such as mathematical expressions. In this paper, we take unit conversions as an example and propose a data augmentation technique which lead to models learning both translation and conversion tasks as well as how to adequately switch between them for end-to-end localization.
Though people rarely speak in complete sentences, punctuation confers many benefits to the readers of transcribed speech. Unfortunately, most ASR systems do not produce punctuated output. To address this, we propose a solution for automatic punctuation that is both cost efficient and easy to train. Our solution benefits from the recent trend in fine-tuning transformer-based language models. We also modify the typical framing of this task by predicting punctuation for sequences rather than individual tokens, which makes for more efficient training and inference. Finally, we find that aggregating predictions across multiple context windows improves accuracy even further. Our best model achieves a new state of the art on benchmark data (TED Talks) with a combined F1 of 83.9, representing a 48.7% relative improvement (15.3 absolute) over the previous state of the art.
Translationese is a phenomenon present in human translations, simultaneous interpreting, and even machine translations. Some translationese features tend to appear in simultaneous interpreting with higher frequency than in human text translation, but the reasons for this are unclear. This study analyzes translationese patterns in translation, interpreting, and machine translation outputs in order to explore possible reasons. In our analysis we – (i) detail two non-invasive ways of detecting translationese and (ii) compare translationese across human and machine translations from text and speech. We find that machine translation shows traces of translationese, but does not reproduce the patterns found in human translation, offering support to the hypothesis that such patterns are due to the model (human vs machine) rather than to the data (written vs spoken).
To empower end users in searching for historical linguistic content with a performance that far exceeds the research functions offered by websites of, e.g., historical dictionaries, is undoubtedly a major advantage of (Linguistic) Linked Open Data ([L]LOD). An important aim of lexicography is to enable a language-independent, onomasiological approach, and the modelling of linguistic resources following the LOD paradigm facilitates the semantic mapping to ontologies making this approach possible. Hallig-Wartburg’s Begriffssystem (HW) is a well-known extra-linguistic conceptual system used as an onomasiological framework by many historical lexicographical and lexicological works. Published in 1952, HW has meanwhile been digitised. With proprietary XML data as the starting point, our goal is the transformation of HW into Linked Open Data in order to facilitate its use by linguistic resources modelled as LOD. In this paper, we describe the particularities of the HW conceptual model and the method of converting HW: We discuss two approaches, (i) the representation of HW in RDF using SKOS, the SKOS thesaurus extension, and XKOS, and (ii) the creation of a lightweight ontology expressed in OWL, based on the RDF/SKOS model. The outcome is illustrated with use cases of medieval Gascon, and Italian.
The Cologne Digital Sanskrit Dictionaries (CDSD) is a large collection of complex digitized Sanskrit dictionaries, consisting of over thirty-five works, and is the most prominent collection of Sanskrit dictionaries worldwide. In this paper we evaluate two methods for transforming the CDSD into Ontolex-Lemon based on a modelling exercise. The first method that we evaluate consists of applying RDFa to the existent TEI-P5 files. The second method consists of transforming the TEI-encoded dictionaries into new files containing RDF triples modelled in OntoLex-Lemon. As a result of the modelling exercise we choose the second method: to transform TEI-encoded lexical data into Ontolex-Lemon by creating new files containing exclusively RDF triples.
The increasing recognition of the utility of Linked Data as a means of publishing lexical resource has helped to underline the need for RDF based data models which have the flexibility and expressivity to be able to represent the most salient kinds of information contained in such resources as structured data, including, notably, information relating to time and the temporal dimension. In this article we describe a perdurantist approach to modelling diachronic lexical information which builds upon work which we have previously presented and which is based on the ontolex-lemon vocabulary. We present two extended examples, one taken from the Oxford English Dictionary, the other from a work on etymology, to show how our approach can handle different kinds of temporal information often found in lexical resources.
Language catalogues and typological databases are two important types of resources containing different types of knowledge about the world’s natural languages. The former provide metadata such as number of speakers, location (in prose descriptions and/or GPS coordinates), language code, literacy, etc., while the latter contain information about a set of structural and functional attributes of languages. Given that both types of resources are developed and later maintained manually, there are practical limits as to the number of languages and the number of features that can be surveyed. We introduce the concept of a language profile, which is intended to be a structured representation of various types of knowledge about a natural language extracted semi-automatically from descriptive documents and stored at a central location. It has three major parts: (1) an introductory; (2) an attributive; and (3) a reference part, each containing different types of knowledge about a given natural language. As a case study, we develop and present a language profile of an example language. At this stage, a language profile is an independent entity, but in the future it is envisioned to become part of a network of language profiles connected to each other via various types of relations. Such a representation is expected to be suitable both for humans and machines to read and process for further deeper linguistic analyses and/or comparisons.
In recent years, there has been increasing interest in publishing lexicographic and terminological resources as linked data. The benefit of using linked data technologies to publish terminologies is that terminologies can be linked to each other, thus creating a cloud of linked terminologies that cross domains, languages and that support advanced applications that do not work with single terminologies but can exploit multiple terminologies seamlessly. We present Terme-‘a-LLOD (TAL), a new paradigm for transforming and publishing terminologies as linked data which relies on a virtualization approach. The approach rests on a preconfigured virtual image of a server that can be downloaded and installed. We describe our approach to simplifying the transformation and hosting of terminological resources in the remainder of this paper. We provide a proof-of-concept for this paradigm showing how to apply it to the conversion of the well-known IATE terminology as well as to various smaller terminologies. Further, we discuss how the implementation of our paradigm can be integrated into existing NLP service infrastructures that rely on virtualization technology. While we apply this paradigm to the transformation and hosting of terminologies as linked data, the paradigm can be applied to any other resource format as well.
We introduce a new dataset for the Linguistic Linked Open Data (LLOD) cloud that will provide metadata about annotation and language information harvested from annotated language resources like corpora freely available on the internet. To our knowledge annotation metadata is not provided by any metadata provider, e.g. linghub, datahub or CLARIN so far. On the other hand, language metadata that is found on such portals is rarely provided in machine-readable form, especially as Linked Data. In this paper, we describe the harvesting process, content and structure of the new dataset and its application in the Lin|gu|is|tik portal, a research platform for linguists. Aside from that, we introduce tools for the conversion of XML encoded language resources to the CoNLL format. The generated RDF data as well as the XML-converter application are made public under an open license.
This paper reports on an ongoing task of monolingual word sense alignment in which a comparative study between the Portuguese Academy of Sciences Dictionary and the Dicionário Aberto is carried out in the context of the ELEXIS (European Lexicographic Infrastructure) project. Word sense alignment involves searching for matching senses within dictionary entries of different lexical resources and linking them, which poses significant challenges. The lexicographic criteria are not always entirely consistent within individual dictionaries and even less so across different projects where different options may have been assumed in terms of structure and especially wording techniques of lexicographic glosses. This hinders the task of matching senses. We aim to present our annotation workflow in Portuguese using the Semantic Web technologies. The results obtained are useful for the discussion within the community.
A practical alignment service should be flexible enough to handle the varied alignment scenarios that arise in the real world, while minimizing the need for manual configuration. MAPLE, an orchestration framework for ontology alignment, supports this goal by coordinating a few loosely coupled actors, which communicate and cooperate to solve a matching task using explicit metadata about the input ontologies, other available resources and the task itself. The alignment task is thus summarized by a report listing its characteristics and suggesting alignment strategies. The schema of the report is based on several metadata vocabularies, among which the Lime module of the OntoLex-Lemon model is particularly important, summarizing the lexical content of the input ontologies and describing external language resources that may be exploited for performing the alignment. In this paper, we propose a REST API that enables the participation of downstream alignment services in the process orchestrated by MAPLE, helping them self-adapt in order to handle heterogeneous alignment tasks and scenarios. The realization of this alignment orchestration effort has been performed through two main phases: we first described its API as an OpenAPI specification (a la API-first), which we then exploited to generate server stubs and compliant client libraries. Finally, we switched our focus to the integration of existing alignment systems, with one fully integrated system and an additional one being worked on, in the effort to propose the API as a valuable addendum to any system being developed.
This paper describes the ongoing work in converting the lexicographic collection of a non-standard German language dataset (Bavarian Dialects) into a Linguistic Linked Open Data (LLOD) format. The collection is divided into two: questionnaire dataset (DBÖ) which contains details of the questionnaires, questions, collectors, paper slips etc., and the lexical dataset (WBÖ) which contains lexical entries (answers) collected in response to the questions. In its current form, the lexical dataset is available in a TEI/XML format separately from the questionnaire dataset. This paper presents the mapping of the lexical entries in the TEI/XML format into LLOD format using the Ontolex-Lemon model. The paper shows how the data in TEI/XML format is transformed into LLOD and produces a lexicon for Bavarian Dialects. It also presents the approach used to interlink the original questions with the lexical entries. The resulting lexicon complements the questionnaire dataset, which is already in a LLOD format, by semantically interlinking the original questions with the answers and vice-versa.
In this contribution, we show LexO, a user-friendly web collaborative editor of lexical resources based on the lemon model. LexO has been developed in the context of Digital Humanities projects, in which a key point in the design of an editor was the ease of use by lexicographers with no skill in Linked Data or Semantic Web technologies. Though the tool already allows creating a lemon lexicon from scratch and lets a team of users work on it collaboratively, many developments are possible. The involvement of the LLOD community appears now crucial both to find new users and application fields where to test it, and, even more importantly, to understand in which way it should evolve.
This paper addresses the task of supervised hypernymy detection in Spanish through an order embedding and using pretrained word vectors as input. Although the task has been widely addressed in English, there is not much work in Spanish, and according to our knowledge there is not any available dataset for supervised hypernymy detection in Spanish. We built a supervised hypernymy dataset for Spanish from WordNet and corpus statistics information, with different versions according to the lexical intersection between its partitions: random and lexical split. We show the results of using the resulting dataset within an order embedding consuming pretrained word vectors as input. We show the ability of pretrained word vectors to transfer learning to unseen lexical units according to the results in the lexical split dataset. To finish, we study the results of giving additional information in training time, such as, cohyponym links and instances extracted through patterns.
Wikidata now records data about lexemes, senses and lexical forms and exposes them as Linguistic Linked Open Data. Since lexemes in Wikidata was first established in 2018, this data has grown considerable in size. Links between lexemes in different languages can be made, e.g., through a derivation property or senses. We present some descriptive statistics about the lexemes of Wikidata, focusing on the multilingual aspects and show that there are still relatively few multilingual links.
Word embeddings such as Word2Vec not only uniquely identify words but also encode important semantic information about them. However, as single entities they are difficult to interpret and their individual dimensions do not have obvious meanings. A more intuitive and interpretable feature space based on neural representations of words was presented by Binder and colleagues (2016) but is only available for a very limited vocabulary. Previous research (Utsumi, 2018) indicates that Binder features can be predicted for words from their embedding vectors (such as Word2Vec), but only looked at the original Binder vocabulary. This paper aimed to demonstrate that Binder features can effectively be predicted for a large number of new words and that the predicted values are sensible. The results supported this, showing that correlations between predicted feature values were consistent with those in the original Binder dataset. Additionally, vectors of predicted values performed comparatively to established embedding models in tests of word-pair semantic similarity. Being able to predict Binder feature space vectors for any number of new words opens up many uses not possible with the original vocabulary size.
Texts comprise a large part of visual information that we process every day, so one of the tasks of language science is to make them more accessible. However, often the text design process is focused on the font size, but not on its type; which might be crucial especially for the people with reading disabilities. The current paper represents a study on text accessibility and the first attempt to create a research-based accessible font for Cyrillic letters. This resulted in the dyslexic-specific font, LexiaD. Its design rests on the reduction of inter-letter similarity of the Russian alphabet. In evaluation stage, dyslexic and non-dyslexic children were asked to read sentences from the Children version of the Russian Sentence Corpus. We tested the readability of LexiaD compared to PT Sans and PT Serif fonts. The results showed that all children had some advantage in letter feature extraction and information integration while reading in LexiaD, but lexical access was improved when sentences were rendered in PT Sans or PT Serif. Therefore, in several aspects, LexiaD proved to be faster to read and could be recommended to use by dyslexics who have visual deficiency or those who struggle with text understanding resulting in re-reading.
NLP models are imperfect and lack intricate capabilities that humans access automatically when processing speech or reading a text. Human language processing data can be leveraged to increase the performance of models and to pursue explanatory research for a better understanding of the differences between human and machine language processing. We review recent studies leveraging different types of cognitive processing signals, namely eye-tracking, M/EEG and fMRI data recorded during language understanding. We discuss the role of cognitive data for machine learning-based NLP methods and identify fundamental challenges for processing pipelines. Finally, we propose practical strategies for using these types of cognitive signals to enhance NLP models.
Linguistics predictability is the degree of confidence in which language unit (word, part of speech, etc.) will be the next in the sequence. Experiments have shown that the correct prediction simplifies the perception of a language unit and its integration into the context. As a result of an incorrect prediction, language processing slows down. Currently, to get a measure of the language unit predictability, a neurolinguistic experiment known as a cloze task has to be conducted on a large number of participants. Cloze tasks are resource-consuming and are criticized by some researchers as an insufficiently valid measure of predictability. In this paper, we compare different language models that attempt to simulate human respondents’ performance on the cloze task. Using a language model to create cloze task simulations would require significantly less time and conduct studies related to linguistic predictability.
While there is a rich literature on the tracking of sentiment and emotion in texts, modelling the emotional trajectory of longer narratives, such as literary texts, poses new challenges. Previous work in the area of sentiment analysis has focused on using information from within a sentence to predict a valence value for that sentence. We propose to explore the influence of previous sentences on the sentiment of a given sentence. In particular, we investigate whether information present in a history of previous sentences can be used to predict a valence value for the following sentence. We explored both linear and non-linear models applied with a range of different feature combinations. We also looked at different context history sizes to determine what range of previous sentence context was the most informative for our models. We establish a linear relationship between sentence context history and the valence value of the current sentence and demonstrate that sentences in closer proximity to the target sentence are more informative. We show that the inclusion of semantic word embeddings further enriches our model predictions.
We present the Le Petit Prince Corpus (LPPC), a multi-lingual resource for research in (computational) psycho- and neurolinguistics. The corpus consists of the children’s story The Little Prince in 26 languages. The dataset is in the process of being built using state-of-the-art methods for speech and language processing and electroencephalography (EEG). The planned release of LPPC dataset will include raw text annotated with dependency graphs in the Universal Dependencies standard, a near-natural-sounding synthetic spoken subset as well as EEG recordings. We will use this corpus for conducting neurolinguistic studies that generalize across a wide range of languages, overcoming typological constraints to traditional approaches. The planned release of the LPPC combines linguistic and EEG data for many languages using fully automatic methods, and thus constitutes a readily extendable resource that supports cross-linguistic and cross-disciplinary research.
In sentiment analysis, several researchers have used emoji and hashtags as specific forms of training and supervision. Some emotions, such as fear and disgust, are underrepresented in the text of social media. Others, such as anticipation, are absent. This research paper proposes a new dataset for complex emotion detection using a combination of several existing corpora in order to represent and interpret complex emotions based on the Plutchik’s theory. Our experiments and evaluations confirm that using Transfer Learning (TL) with a rich emotional corpus, facilitates the detection of complex emotions in a four-dimensional space. In addition, the incorporation of the rule on the reverse emotions in the model’s architecture brings a significant improvement in terms of precision, recall, and F-score.
Embodied cognitive science suggested a number of variables describing our sensorimotor experience associated with different concepts: modality experience rating (i.e., relationship between words and images of a particular perceptive modality—visual, auditory, haptic etc.), manipulability (the necessity for an object to interact with human hands in order to perform its function), vertical spatial localization. According to the embodied cognition theory, these semantic variables capture our mental representations and thus should influence word learning, processing and production. However, it is not clear how these new variables are related to such traditional variables as imageability, age of acquisition (AoA) and word frequency. In the presented database, normative data on the modality (visual, auditory, haptic, olfactory, and gustatory) ratings, vertical spatial localization of the object, manipulability, imageability, age of acquisition, and subjective frequency for 506 Russian nouns are collected. Factor analysis revealed four factors: (1) visual and haptic modality ratings were combined with imageability, manipulability and AoA; (2) word length, frequency and AoA; (3) olfactory modality was united with gustatory; (4) spatial localization only was included in the fourth factor. The database is available online together with a publication describing the method of data collection and data parameters (Miklashevsky, 2018).
Today scientific workflows are used by scientists as a way to define automated, scalable, and portable in-silico experiments. Having a formal description of an experiment can improve replicability and reproducibility of the experiment. However, simply storing and publishing the workflow may be not enough, an accurate management of provenance data generated during workflow life cycle is crucial to achieve reproducibility. This document presents the activity being carried out by CNR-ISTI in task 5.2 of the SSHOC project to add to the repository service developed in the task, functionalities to store, access and manage ‘workflow data’ in order to improve replicability and reproducibility of e-science experiments.
The paper presents a journey, which starts from various social sciences and humanities (SSH) Research Infrastructures in Europe and arrives at the comprehensive “ecosystem of infrastructures”, namely the European Open Science Cloud (EOSC). We will highlight how the SSH Open Science infrastructures contribute to the goal of establishing the EOSC. First, through the example of OPERAS, the European Research Infrastructure for Open Scholarly Communication in the SSH, to see how its services are conceived to be part of the EOSC and to address the communities’ needs. The next two sections highlight collaboration practices between partners in Europe to build the SSH component of the EOSC and a SSH discovery platform, as a service of OPERAS and the EOSC. The last two sections will focus on an implementation network dedicated to SSH data fairification.
This paper describes a collection of 20k ELAN annotation files harvested from five different endangered language archives. The ELAN files form a very heterogeneous set, but the hierarchical configuration of their tiers allow, in conjunction with the tier content, to identify transcriptions, translations, and glosses. These transcriptions, translations, and glosses are queryable across archives. Small analyses of graphemes (transcription tier), grammatical and lexical glosses (gloss tier), and semantic concepts (translation tier) show the viability of the approach. The use of identifiers from OLAC, Wikidata and Glottolog allows for a better integration of the data from these archives into the Linguistic Linked Open Data Cloud.
We present a replication of a data-driven and linguistically inspired Verbal Aggression analysis framework that was designed to examine Twitter verbal attacks against predefined target groups of interest as an indicator of xenophobic attitudes during the financial crisis in Greece, in particular during the period 2013-2016. The research goal in this paper is to re-examine Verbal Aggression as an indicator of xenophobic attitudes in Greek Twitter three years later, in order to trace possible changes regarding the main targets, the types and the content of the verbal attacks against the same targets in the post crisis era, given also the ongoing refugee crisis and the political landscape in Greece as it was shaped after the elections in 2019. The results indicate an interesting rearrangement of the main targets of the verbal attacks, while the content and the types of the attacks provide valuable insights about the way these targets are being framed as compared to the respective dominant perceptions and stereotypes about them during the period 2013-2016.
For the analysis of historical wage development, no structured data is available. Job advertisements, as found in newspapers can provide insights into what different types of jobs paid, but require language technology to structure in a format conducive to quantitative analysis. In this paper, we report on our experiments to mine wages from 19th century newspaper advertisements and detail the challenges that need to be overcome to perform a socio-economic analysis of textual data sources.
This paper outlines the future of language resources and identifies their potential contribution for creating and sustaining the social sciences and humanities (SSH) component of the European Open Science Cloud (EOSC).
This paper aims to give some insights on how the European Open Science Cloud (EOSC) will be able to influence the Social Sciences and Humanities (SSH) sector, thus paving the way towards innovation. Points of discussion on how the LRs and RIs community can contribute to the revolution in the practice of research areas are provided.
Given the persistent gap between demand and supply, the impetus to reuse language resources is great. Researchers benefit from building upon the work of others including reusing data, tools and methodology. Such reuse should always consider the original intent of the language resource and how that impacts potential reanalysis. When the reuse crosses disciplinary boundaries, the re-user also needs to consider how research standards that differ between social science and humanities on the one hand and human language technologies on the other might lead to differences in unspoken assumptions. Data centers that aim to support multiple research communities have a responsibility to build bridges across disciplinary divides by sharing data in all directions, encouraging re-use and re-sharing and engaging directly in research that improves methodologies.
Spoken audio data, such as interview data, is a scientific instrument used by researchers in various disciplines crossing the boundaries of social sciences and humanities. In this paper, we will have a closer look at a portal designed to perform speech-to-text conversion on audio recordings through Automatic Speech Recognition (ASR) in the CLARIN infrastructure. Within the cluster cross-domain EU project SSHOC the potential value of such a linguistic tool kit for processing spoken language recording has found uptake in a webinar about the topic, and in a task addressing audio analysis of panel survey data. The objective of this contribution is to show that the processing of interviews as a research instrument has opened up a fascinating and fruitful area of collaboration between Social Sciences and Humanities (SSH).
The disability benefits programs administered by the US Social Security Administration (SSA) receive between 2 and 3 million new applications each year. Adjudicators manually review hundreds of evidence pages per case to determine eligibility based on financial, medical, and functional criteria. Natural Language Processing (NLP) technology is uniquely suited to support this adjudication work and is a critical component of an ongoing inter-agency collaboration between SSA and the National Institutes of Health. This NLP work provides resources and models for document ranking, named entity recognition, and terminology extraction in order to automatically identify documents and reports pertinent to a case, and to allow adjudicators to search for and locate desired information quickly. In this paper, we describe our vision for how NLP can impact SSA’s adjudication process, present the resources and models that have been developed, and discuss some of the benefits and challenges in working with large-scale government data, and its specific properties in the functional domain.
In this paper, we propose FRAQUE, a question answering system for factoid questions in the Public administration domain. The system is based on semantic frames, here intended as collections of slots typed with their possible values. FRAQUE queries unstructured textual data and exploits the potential of different approaches: it extracts pattern elements from texts which are linguistically analyzed through statistical methods.FRAQUE allows Italian users to query vast document repositories related to the domain of Public Administration. Given the statistical nature of most of its components such as word embeddings, the system allows for a flexible domain and language adaptation process. FRAQUE’s goal is to associate questions with frames stored into a Knowledge Graph along with relevant document passages, which are returned as the answer.
In this paper, we show the enhancing of the Demanded Skills Diagnosis (DiCoDe: Diagnóstico de Competencias Demandadas), a system developed by Mexico City’s Ministry of Labor and Employment Promotion (STyFE: Secretaría de Trabajo y Fomento del Empleo de la Ciudad de México) that seeks to reduce information asymmetries between job seekers and employers. The project uses webscraping techniques to retrieve job vacancies posted on private job portals on a daily basis and with the purpose of informing training and individual case management policies as well as labor market monitoring. For this purpose, a collaboration project between STyFE and the Language Engineering Group (GIL: Grupo de Ingeniería Lingüística) was established in order to enhance DiCoDe by applying NLP models and semantic analysis. By this collaboration, DiCoDe’s job vacancies system’s macro-structure and its geographic referencing at the city hall (municipality) level were improved. More specifically, dictionaries were created to identify demanded competencies, skills and abilities (CSA) and algorithms were developed for dynamic classifying of vacancies and identifying terms for searches on free text, in order to improve the results and processing time of queries.
Cat. 2 Show-case: We present the Data4Impact (D4I) platform, a novel end-to-end system for evidence-based, timely and accurate monitoring and evaluation of research and innovation (R&I) activities. Using the latest technological advances in Human Language Technology (HLT) and our data-driven methodology, we build a novel set of indicators in order to track funded projects and their impact on science, the economy and the society as a whole, during and after the project life-cycle. We develop our methodology by targeting Health-related EC projects from 2007 to 2019 to produce solutions that meet the needs of stakeholders (mainly policy-makers and research funders). Various D4I text analytics workflows process datasets and their metadata, extract valuable insights and estimate intermediate results and metrics, culminating in a set of robust indicators that the users can interact with through our dashboard, the D4I Monitor (available at monitor.data4impact.eu). Therefore, our approach, which can be generalized to different contexts, is multidimensional (technology, tools, indicators, dashboard) and the resulting system can provide an innovative solution for public administrators in their policy-making needs related to RDI funding allocation.
The Austrian Language Resource Portal (Sprachressourcenportal Österreichs) is Austria’s central platform for language resources in the area of public administration. It focuses on language resources in the Austrian variety of the German language. As a product of the cooperation between a public administration body and a university, the Portal contains various language resources (terminological resources in the public administration domain, a language guide, named entities based on open public data, translation memories, etc.). German is a pluricentric language that considerably varies in the domain of public administration due to different public administration systems. Therefore, the Austrian Language Resource Portal stresses the importance of language resources specific to a language variety, thus paving the way for the re-use of variety-specific language data for human language technology, such as machine translation training, for the Austrian standard variety.
Legal-ES is an open source resource kit for legal Spanish. It consists of a large scale Spanish corpus of open legal texts and different kinds of language models including word embeddings and topic models. The corpus includes over 1000 million words covering a collection of legislative and administrative open access documents in Spanish from different sources representing international, national and regional entities. The corpus is pre-processed and tokenized using Spacy. For the word embeddings, gensim was used on the collection of tokens, producing a representation space that is especially suited to reflect the inherent characteristics of the legal domain. We calculate also topic models to obtain a convenient tool to understand the main topics in the corpus and to navigate through the documents exploiting the semantic similarity among documents. We will analyse the time structure of a dynamic topic model to infer changes in the legal production of Spanish jurisdiction that have occurred over the analysed time framework.
This paper introduces and evaluates a Bayesian mixture model that is designed for dating texts based on the distributions of linguistic features. The model is applied to the corpus of Vedic Sanskrit the historical structure of which is still unclear in many details. The evaluation concentrates on the interaction between time, genre and linguistic features, detecting those whose distributions are clearly coupled with the historical time. The evaluation also highlights the problems that arise when quantitative results need to be reconciled with philological insights.
Aramaic is an ancient Semitic language with a 3,000 year history. However, since the number of Aramaic speakers in the world hasdeclined, Aramaic is in danger of extinction. In this paper, we suggest a methodology for automatic construction of Aramaic-Hebrew translation Lexicon. First, we generate an initial translation lexicon by a state-of-the-art word alignment translation model. Then,we filter the initial lexicon using string similarity measures of three types: similarity between terms in the target language, similarity between a source and a target term, and similarity between terms in the source language. In our experiments, we use a parallel corporaof Biblical Aramaic-Hebrew sentence pairs and evaluate various string similarity measures for each type of similarity. We illustratethe empirical benefit of our methodology and its effect on precision and F1. In particular, we demonstrate that our filtering methodsignificantly exceeds a filtering approach based on the probability scores given by a state-of-the-art word alignment translation model.
Automatic dating of ancient documents is a very important area of research for digital humanities applications. Many documents available via digital libraries do not have any dating or dating that is uncertain. Document dating is not only useful by itself but it also helps to choose the appropriate NLP tools (lemmatizer, POS tagger ) for subsequent analysis. This paper provides a dataset with thousands of ancient documents in French and present methods and evaluation metrics for this task. We compare character-level methods with token-level methods on two different datasets of two different time periods and two different text genres. Our results show that character-level models are more robust to noise than classical token-level models. The experiments presented in this article focused on documents written in French but we believe that the ability of character-level models to handle noise properly would help to achieve comparable results on other languages and more ancient languages in particular.
Classical Armenian, Old Georgian and Syriac are under-resourced digital languages. Even though a lot of printed critical editions or dictionaries are available, there is currently a lack of fully tagged corpora that could be reused for automatic text analysis. In this paper, we introduce an ongoing project of lemmatization and POS-tagging for these languages, relying on a recurrent neural network (RNN), specific morphological tags and dedicated datasets. For this paper, we have combine different corpora previously processed by automatic out-of-context lemmatization and POS-tagging, and manual proofreading by the collaborators of the GREgORI Project (UCLouvain, Louvain-la-Neuve, Belgium). We intend to compare a rule based approach and a RNN approach by using PIE specialized by Calfa (Paris, France). We introduce here first results. We reach a mean accuracy of 91,63% in lemmatization and of 92,56% in POS-tagging. The datasets, which were constituted and used for this project, are not yet representative of the different variations of these languages through centuries, but they are homogenous and allow reaching tangible results, paving the way for further analysis of wider corpora.
Traditionally, historical phonologists have relied on tedious manual derivations to calibrate the sequences of sound changes that shaped the phonological evolution of languages. However, humans are prone to errors, and cannot track thousands of parallel word derivations in any efficient manner. We propose to instead automatically derive each lexical item in parallel, and we demonstrate forward reconstruction as both a computational task with metrics to optimize, and as an empirical tool for inquiry. For this end we present DiaSim, a user-facing application that simulates “cascades” of diachronic developments over a language’s lexicon and provides diagnostics for “debugging” those cascades. We test our methodology on a Latin-to-French reflex prediction task, using a newly compiled dataset FLLex with 1368 paired Latin/French forms. We also present, FLLAPS, which maps 310 Latin reflexes through five stages until Modern French, derived from Pope (1934)’s sound tables. Our publicly available rule cascades include the baselines BaseCLEF and BaseCLEF*, representing the received view of Latin to French development, and DiaCLEF, build by incremental corrections to BaseCLEF aided by DiaSim’s diagnostics. DiaCLEF vastly outperforms the baselines, improving final accuracy on FLLex from 3.2%to 84.9%, and similar improvements across FLLAPS’ stages.
This paper presents LatInfLexi, a large inflected lexicon of Latin providing information on all the inflected wordforms of 3,348 verbs and 1,038 nouns. After a description of the structure of the resource and some data on its size, the procedure followed to obtain the lexicon from the database of the Lemlat 3.0 morphological analyzer is detailed, as well as the choices made regarding overabundant and defective cells. The way in which the data of LatInfLexi can be exploited in order to perform a quantitative assessment of predictability in Latin verb inflection is then illustrated: results obtained by computing the conditional entropy of guessing the content of a paradigm cell assuming knowledge of one wordform or multiple wordforms are presented in turn, highlighting the descriptive and theoretical relevance of the analysis. Lastly, the paper envisages the advantages of an inclusion of LatInfLexi into the LiLa knowledge base, both for the presented resource and for the knowledge base itself.
Optical character recognition (OCR) for historical documents is a complex procedure subject to a unique set of material issues, including inconsistencies in typefaces and low quality scanning. Consequently, even the most sophisticated OCR engines produce errors. This paper reports on a tool built for postediting the output of Tesseract, more specifically for correcting common errors in digitized historical documents. The proposed tool suggests alternatives for word forms not found in a specified vocabulary. The assumed error is replaced by a presumably correct alternative in the post-edition based on the scores of a Language Model (LM). The tool is tested on a chapter of the book An Essay Towards Regulating the Trade and Employing the Poor of this Kingdom (Cary, 1719). As demonstrated below, the tool is successful in correcting a number of common errors. If sometimes unreliable, it is also transparent and subject to human intervention.
The basic tasks of ancient Chinese information processing include automatic sentence segmentation, word segmentation, part-of-speech tagging and named entity recognition. Tasks such as lexical analysis need to be based on sentence segmentation because of the reason that a plenty of ancient books are not punctuated. However, step-by-step processing is prone to cause multi-level diffusion of errors. This paper designs and implements an integrated annotation system of sentence segmentation and lexical analysis. The BiLSTM-CRF neural network model is used to verify the generalization ability and the effect of sentence segmentation and lexical analysis on different label levels on four cross-age test sets. Research shows that the integration method adopted in ancient Chinese improves the F1-score of sentence segmentation, word segmentation and part of speech tagging. Based on the experimental results of each test set, the F1-score of sentence segmentation reached 78.95, with an average increase of 3.5%; the F1-score of word segmentation reached 85.73%, with an average increase of 0.18%; and the F1-score of part-of-speech tagging reached 72.65, with an average increase of 0.35%.
This paper describes a first attempt to automatic semantic role labeling in Ancient Greek, using a supervised machine learning approach. A Random Forest classifier is trained on a small semantically annotated corpus of Ancient Greek, annotated with a large amount of linguistic features, including form of the construction, morphology, part-of-speech, lemmas, animacy, syntax and distributional vectors of Greek words. These vectors turned out to be more important in the model than any other features, likely because they are well suited to handle a low amount of training examples. Overall labeling accuracy was 0.757, with large differences with respect to the specific role that was labeled and with respect to text genre. Some ways to further improve these results include expanding the amount of training examples, improving the quality of the distributional vectors and increasing the consistency of the syntactic annotation.
We built a thesaurus for Biblical Hebrew, with connections between roots based on phonetic, semantic, and distributional similarity. To this end, we apply established algorithms to find connections between headwords based on existing lexicons and other digital resources. For semantic similarity, we utilize the cosine-similarity of tf-idf vectors of English gloss text of Hebrew headwords from Ernest Klein’s A Comprehensive Etymological Dictionary of the Hebrew Language for Readers of English as well as to Brown-Driver-Brigg’s Hebrew Lexicon. For phonetic similarity, we digitize part of Matityahu Clark’s Etymological Dictionary of Biblical Hebrew, grouping Hebrew roots into phonemic classes, and establish phonetic relationships between headwords in Klein’s Dictionary. For distributional similarity, we consider the cosine similarity of PPMI vectors of Hebrew roots and also, in a somewhat novel approach, apply Word2Vec to a Biblical corpus reduced to its lexemes. The resulting resource is helpful to those trying to understand Biblical Hebrew, and also stands as a good basis for programs trying to process the Biblical text.
The Voynich Manuscript has baffled scholars for centuries. Some believe the elaborate 15th century codex to be a hoax whilst others believe it is a real medieval manuscript whose contents are as yet unknown. In this paper, we provide additional evidence that the text of the manuscript displays the hallmarks of a proper natural language with respect to the relationship between word probabilities and (i) average information per subword segment and (ii) the relative positioning of consecutive subword segments necessary to uniquely identify words of different probabilities.
Cognate prediction and proto-form reconstruction are key tasks in computational historical linguistics that rely on the study of sound change regularity. Solving these tasks appears to be very similar to machine translation, though methods from that field have barely been applied to historical linguistics. Therefore, in this paper, we investigate the learnability of sound correspondences between a proto-language and daughter languages for two machine-translation-inspired models, one statistical, the other neural. We first carry out our experiments on plausible artificial languages, without noise, in order to study the role of each parameter on the algorithms respective performance under almost perfect conditions. We then study real languages, namely Latin, Italian and Spanish, to see if those performances generalise well. We show that both model types manage to learn sound changes despite data scarcity, although the best performing model type depends on several parameters such as the size of the training data, the ambiguity, and the prediction direction.
We address the problem of creating and evaluating quality Neo-Latin word embeddings for the purpose of philosophical research, adapting the Nonce2Vec tool to learn embeddings from Neo-Latin sentences. This distributional semantic modeling tool can learn from tiny data incrementally, using a larger background corpus for initialization. We conduct two evaluation tasks: definitional learning of Latin Wikipedia terms, and learning consistent embeddings from 18th century Neo-Latin sentences pertaining to the concept of mathematical method. Our results show that consistent Neo-Latin word embeddings can be learned from this type of data. While our evaluation results are promising, they do not reveal to what extent the learned models match domain expert knowledge of our Neo-Latin texts. Therefore, we propose an additional evaluation method, grounded in expert-annotated data, that would assess whether learned representations are conceptually sound in relation to the domain of study.
Although there are several sources where to find historical texts, they usually are available in the original language that makes them generally inaccessible. This paper presents the development of state-of-the-art Neural Machine Systems for the low-resourced Latin-Spanish language pair. First, we build a Transformer-based Machine Translation system on the Bible parallel corpus. Then, we build a comparable corpus from Saint Augustine texts and their translations. We use this corpus to study the domain adaptation case from the Bible texts to Saint Augustine’s works. Results show the difficulties of handling a low-resourced language as Latin. First, we noticed the importance of having enough data, since the systems do not achieve high BLEU scores. Regarding domain adaptation, results show how using in-domain data helps systems to achieve a better quality translation. Also, we observed that it is needed a higher amount of data to perform an effective vocabulary extension that includes in-domain vocabulary.
Fictional prose can be broadly divided into narrative and discursive forms with direct speech being central to any discourse representation (alongside indirect reported speech and free indirect discourse). This distinction is crucial in digital literary studies and enables interesting forms of narratological or stylistic analysis. The difficulty of automatically detecting direct speech, however, is currently under-estimated. Rule-based systems that work reasonably well for modern languages struggle with (the lack of) typographical conventions in 19th-century literature. While machine learning approaches to sequence modeling can be applied to solve the task, they typically face a severed skewness in the availability of training material, especially for lesser resourced languages. In this paper, we report the result of a multilingual approach to direct speech detection in a diverse corpus of 19th-century fiction in 9 European languages. The proposed method finetunes a transformer architecture with multilingual sentence embedder on a minimal amount of annotated training in each language, and improves performance across languages with ambiguous direct speech marking, in comparison to a carefully constructed regular expression baseline.
This paper describes the first edition of EvaLatin, a campaign totally devoted to the evaluation of NLP tools for Latin. The two shared tasks proposed in EvaLatin 2020, i. e. Lemmatization and Part-of-Speech tagging, are aimed at fostering research in the field of language technologies for Classical languages. The shared dataset consists of texts taken from the Perseus Digital Library, processed with UDPipe models and then manually corrected by Latin experts. The training set includes only prose texts by Classical authors. The test set, alongside with prose texts by the same authors represented in the training set, also includes data relative to poetry and to the Medieval period. This also allows us to propose the Cross-genre and Cross-time subtasks for each task, in order to evaluate the portability of NLP tools for Latin across different genres and time periods. The results obtained by the participants for each task and subtask are presented and discussed.
Textual data in ancient and historical languages such as Latin is increasingly available in machine readable forms, yet computational tools to analyze and process this data are still lacking. We describe our system for part-of-speech tagging in Latin, an entry in the EvaLatin 2020 shared task. Based on a detailed analysis of the training data, we make targeted preprocessing decisions and design our model. We leverage existing large unlabelled resources to pre-train representations at both the grapheme and word level, which serve as the inputs to our LSTM-based models. We perform an extensive cross-validated hyperparameter search, achieving an accuracy score of up to 93 on in-domain texts. We publicly release all our code and trained models in the hope that our system will be of use to social scientists and digital humanists alike. The insights we draw from our inital analysis can also inform future NLP work modeling syntactic information in Latin.
We describe the JHUBC submission to the EvaLatin Shared task on lemmatization and part-of-speech tagging for Latin. We modify a hard-attentional character-based encoder-decoder to produce lemmas and POS tags with separate decoders, and to incorporate contextual tagging cues. While our results show that the dual decoder approach fails to encode data as successfully as the single encoder, our simple context incorporation method does lead to modest improvements.
The paper presents the system used in the EvaLatin shared task to POS tag and lemmatize Latin. It consists of two components. A gradient boosting machine (LightGBM) is used for POS tagging, mainly fed with pre-computed word embeddings of a window of seven contiguous tokens—the token at hand plus the three preceding and following ones—per target feature value. Word embeddings are trained on the texts of the Perseus Digital Library, Patrologia Latina, and Biblioteca Digitale di Testi Tardo Antichi, which together comprise a high number of texts of different genres from the Classical Age to Late Antiquity. Word forms plus the outputted POS labels are used to feed a seq2seq algorithm implemented in Keras to predict lemmas. The final shared-task accuracies measured for Classical Latin texts are in line with state-of-the-art POS taggers (∼0.96) and lemmatizers (∼0.95).
We present our contribution to the EvaLatin shared task, which is the first evaluation campaign devoted to the evaluation of NLP tools for Latin. We submitted a system based on UDPipe 2.0, one of the winners of the CoNLL 2018 Shared Task, The 2018 Shared Task on Extrinsic Parser Evaluation and SIGMORPHON 2019 Shared Task. Our system places first by a wide margin both in lemmatization and POS tagging in the open modality, where additional supervised data is allowed, in which case we utilize all Universal Dependency Latin treebanks. In the closed modality, where only the EvaLatin training data is allowed, our system achieves the best performance in lemmatization and in classical subtask of POS tagging, while reaching second place in cross-genre and cross-time settings. In the ablation experiments, we also evaluate the influence of BERT and XLM-RoBERTa contextualized embeddings, and the treebank encodings of the different flavors of Latin treebanks.
Despite the great importance of the Latin language in the past, there are relatively few resources available today to develop modern NLP tools for this language. Therefore, the EvaLatin Shared Task for Lemmatization and Part-of-Speech (POS) tagging was published in the LT4HALA workshop. In our work, we dealt with the second EvaLatin task, that is, POS tagging. Since most of the available Latin word embeddings were trained on either few or inaccurate data, we trained several embeddings on better data in the first step. Based on these embeddings, we trained several state-of-the-art taggers and used them as input for an ensemble classifier called LSTMVoter. We were able to achieve the best results for both the cross-genre and the cross-time task (90.64% and 87.00%) without using additional annotated data (closed modality). In the meantime, we further improved the system and achieved even better results (96.91% on classical, 90.87% on cross-genre and 87.35% on cross-time).
Previous studies have shown that the knowledge about attributes and properties in the SUMO ontology and its mapping to WordNet adjectives lacks of an accurate and complete characterization. A proper characterization of this type of knowledge is required to perform formal commonsense reasoning based on the SUMO properties, for instance to distinguish one concept from another based on their properties. In this context, we propose a new semi-automatic approach to model the knowledge about properties and attributes in SUMO by exploiting the information encoded in WordNet adjectives and its mapping to SUMO. To that end, we considered clusters of semantically related groups of WordNet adjectival and nominal synsets. Based on these clusters, we propose a new semi-automatic model for SUMO attributes and their mapping to WordNet, which also includes polarity information. In this paper, as an exploratory approach, we focus on qualities.
Due to rapid urbanization and a homogenized medium of instruction imposed in educational institutions, we have lost much of the golden literary offerings of the diverse languages and dialects that India once possessed. There is an urgent need to mitigate the paucity of online linguistic resources for several Hindi dialects. Given the corpus of a dialect, our system integrates the vocabulary of the dialect to the synsets of IndoWordnet along with their corresponding meta-data. Furthermore, we propose a systematic method for generating exemplary sentences for each newly integrated dialect word. The vocabulary thus integrated follows the schema of the wordnet and generates exemplary sentences to illustrate the meaning and usage of the word. We illustrate our methodology with the integration of words in the Awadhi dialect to the Hindi IndoWordnet to achieve an enrichment of 11.68 % to the existing Hindi synsets. The BLEU metric for evaluating the quality of sentences yielded a 75th percentile score of 0.6351.
WordNet, while one of the most widely used resources for NLP, has not been updated for a long time, and as such a new project English WordNet has arisen to continue the development of the model under an open-source paradigm. In this paper, we detail the second release of this resource entitled “English WordNet 2020”. The work has focused firstly, on the introduction of new synsets and senses and developing guidelines for this and secondly, on the integration of contributions from other projects. We present the changes in this edition, which total over 15,000 changes over the previous release.
Wordnets are lexical databases where the semantic relations of words and concepts are established. These resources are useful for manyNLP tasks, such as automatic text classification, word-sense disambiguation or machine translation. In comparison with other wordnets,the Basque version is smaller and some PoS are underrepresented or missing e.g. adjectives and adverbs. In this work, we explore anovel approach to enrich the Basque WordNet, focusing on the adjectives. We want to prove the use and and effectiveness of sentimentlexicons to enrich the resource without the need of starting from scratch. Using as complementary resources, one dictionary and thesentiment valences of the words, we check if the word of the lexicon matches with the meaning of the synset, and if it matches we addthe word as variant to the Basque WordNet. Following this methodology, we describe the most frequent adjectives with positive andnegative valence, the matches and the possible solutions for the non-matches.
Information systems gathering big amounts of resources growing with time containing distinct modalities (text, audio, video, images, GIS) and aggregating content in various ways (modular e-learning modules, Web systems presenting cultural artefacts) require tools supporting content description. The subject of the description may be the topic and the characteristics of the content expressed by sets of attributes. To describe such resources one can just use some of existing indexing languages like thesauri, classification systems, domain and upper ontologies, terminologies or dictionaries. When appropriate language does not exist, it is necessary to build a new system, which will have to serve both experts who describe resources and non-experts who search through them. The solution presented in this paper used to resource description, allows experts to freely select words and expressions, which are organized in hierarchies of various nature, including that of domain and application character. This is based on the wordnet structure, which introduces a clear order for each of these groups due to its lexical nature. The paper presents two systems where such approach was applied: the E-archaeology.org e-learning content repository in which domain knowledge was integrated to describe content topics and the Hatch system gathering multimodal information about the archaeological site targeted at a wide audience, where application conceptualization was applied to describe the content by a set of attributes.
We extend the Open WordNet for English (OWN-EN) with rock-related and other lithological terms using the authoritative source of GBA’s Thesaurus. Our aim is to improve WordNet to better function within Oil & Gas domain, particularly geoscience texts. We use a three step approach: a proof of concept-level extension of WordNet, a major extension on which we evaluate the impact with positive results and a full extension encompassing all GBA’s lithological terms. We also build a mapping to GBA which also links to several other resources: WikiData, British Geological Survey, Inspire, GeoSciML and DBpedia.
We describe on-going work consisting in adding pronunciation information to wordnets, as such information can indicate specific senses of a word. Many wordnets associate with their senses only a lemma form and a part-of-speech tag. At the same time, we are aware that additional linguistic information can be useful for identifying a specific sense of a wordnet lemma when encountered in a corpus. While work already deals with the addition of grammatical number or grammatical gender information to wordnet lemmas,we are investigating the linking of wordnet lemmas to pronunciation information, adding thus a speech-related modality to wordnets
Electronic Health Records are a valuable source of patient information which can be leveraged to detect Adverse Drug Events (ADEs) and aid post-mark drug-surveillance. The overall aim of this study is to scrutinize text written by clinicians in the EHRs and build a model for ADE detection that produces medically relevant predictions. Natural Language Processing techniques will be exploited to create important predictors and incorporate them into the learning process. The study focuses on the 5 most frequent ADE cases found ina Swedish electronic patient record corpus. The results indicate that considering textual features, rather than the structured, can improve the classification performance by 15% in some ADE cases. Additionally, variable patient history lengths are incorporated in the models, demonstrating the importance of the above decision rather than using an arbitrary number for a history length. The experimental findings suggest that the clinical text in EHRs includes information that can capture data beyond the ones that are found in a structured format.
We present a large Norwegian lexical resource of categorized medical terms. The resource, which merges information from large medical databases, contains over 56,000 entries, including automatically mapped terms from a Norwegian medical dictionary. We describe the methodology behind this automatic dictionary entry mapping based on keywords and suffixes and further present the results of a manual evaluation performed on a subset by a domain expert. The evaluation indicated that ca. 80% of the mappings were correct.
Medical language exhibits great variations regarding users, institutions and language registers. With large parts of clinical documents in free text, NLP is playing a more and more important role in unlocking re-usable and interoperable meaning from medical records. This study describes the architectural principles and the evolution of a German interface vocabulary, combining machine translation with human annotation and rule-based term generation, yielding a resource with 7.7 million raw entries, each of which linked to the reference terminology SNOMED CT, an international standard with about 350 thousand concepts. The purpose is to offer a high coverage of medical jargon in order to optimise terminology grounding of clinical texts by text mining systems. The core resource is a manually curated table of English-to-German word and chunk translations, supported by a set of language generation rules. The work describes a workflow consisting the enrichment and modification of this table with human and machine efforts, manually enriched by grammarspecific tags. Top-down and bottom-up methods for terminology population used in parallel. The final interface terms are produced by a term generator, which creates one-to-many German variants per SNOMED CT English description. Filtering against a large collection of domain terminologies and corpora drastically reduces the size of the vocabulary in favour of more realistic terms or terms that can reasonably be expected to match clinical text passages within a text-mining pipeline. An evaluation was performed by a comparison between the current version of the German interface vocabulary and the English description table of the SNOMED CT International release. An exact term matching was performed with a small parallel corpus constituted by text snippets from different clinical documents. With overall low retrieval parameters (with F-values around 30%), the performance of the German language scenario reaches 80 – 90% of the English one. Interestingly, annotations are slightly better with machine-translated (German – English) texts, using the International SNOMED CT resource only.
Translating biomedical ontologies is an important challenge, but doing it manually requires much time and money. We study the possibility to use open-source knowledge bases to translate biomedical ontologies. We focus on two aspects: coverage and quality. We look at the coverage of two biomedical ontologies focusing on diseases with respect to Wikidata for 9 European languages (Czech, Dutch, English, French, German, Italian, Polish, Portuguese and Spanish) for both, plus Arabic, Chinese and Russian for the second. We first use direct links between Wikidata and the studied ontologies and then use second-order links by going through other intermediate ontologies. We then compare the quality of the translations obtained thanks to Wikidata with a commercial machine translation tool, here Google Cloud Translation.
Pre-trained text encoders have rapidly advanced the state-of-the-art on many Natural Language Processing tasks. This paper presents the use of transfer learning methods applied to the automatic detection of codes in radiological reports in Spanish. Assigning codes to a clinical document is a popular task in NLP and in the biomedical domain. These codes can be of two types: standard classifications (e.g. ICD-10) or specific to each clinic or hospital. In this study we show a system using specific radiology clinic codes. The dataset is composed of 208,167 radiology reports labeled with 89 different codes. The corpus has been evaluated with three methods using the BERT model applied to Spanish: Multilingual BERT, BETO and XLM. The results are interesting obtaining 70% of F1-score with a pre-trained multilingual model.
The Platform for Automated extraction of animal Disease Information from the web (PADI-web) is an automated system which monitors the web for monitoring and detecting emerging animal infectious diseases. The tool automatically collects news via customised multilingual queries, classifies them and extracts epidemiological information. We detail the processing of multilingual online sources by PADI-web and analyse the translated outputs in a case study
We describe the finding of the Fourth Workshop on Neural Generation and Translation, held in concert with the annual conference of the Association for Computational Linguistics (ACL 2020). First, we summarize the research trends of papers presented in the proceedings. Second, we describe the results of the three shared tasks 1) efficient neural machine translation (NMT) where participants were tasked with creating NMT systems that are both accurate and efficient, and 2) document-level generation and translation (DGT) where participants were tasked with developing systems that generate summaries from structured data, potentially with assistance from text in another language and 3) STAPLE task: creation of as many possible translations of a given input text. This last shared task was organised by Duolingo.
Text style transfer refers to the task of rephrasing a given text in a different style. While various methods have been proposed to advance the state of the art, they often assume the transfer output follows a delta distribution, and thus their models cannot generate different style transfer results for a given input text. To address the limitation, we propose a one-to-many text style transfer framework. In contrast to prior works that learn a one-to-one mapping that converts an input sentence to one output sentence, our approach learns a one-to-many mapping that can convert an input sentence to multiple different output sentences, while preserving the input content. This is achieved by applying adversarial training with a latent decomposition scheme. Specifically, we decompose the latent representation of the input sentence to a style code that captures the language style variation and a content code that encodes the language style-independent content. We then combine the content code with the style code for generating a style transfer output. By combining the same content code with a different style code, we generate a different style transfer output. Extensive experimental results with comparisons to several text style transfer approaches on multiple public datasets using a diverse set of performance metrics validate effectiveness of the proposed approach.
We propose a novel procedure for training multiple Transformers with tied parameters which compresses multiple models into one enabling the dynamic choice of the number of encoder and decoder layers during decoding. In training an encoder-decoder model, typically, the output of the last layer of the N-layer encoder is fed to the M-layer decoder, and the output of the last decoder layer is used to compute loss. Instead, our method computes a single loss consisting of NxM losses, where each loss is computed from the output of one of the M decoder layers connected to one of the N encoder layers. Such a model subsumes NxM models with different number of encoder and decoder layers, and can be used for decoding with fewer than the maximum number of encoder and decoder layers. Given our flexible tied model, we also address to a-priori selection of the number of encoder and decoder layers for faster decoding, and explore recurrent stacking of layers and knowledge distillation for model compression. We present a cost-benefit analysis of applying the proposed approaches for neural machine translation and show that they reduce decoding costs while preserving translation quality.
Neural Machine Translation (NMT) is resource-intensive. We design a quantization procedure to compress fit NMT models better for devices with limited hardware capability. We use logarithmic quantization, instead of the more commonly used fixed-point quantization, based on the empirical fact that parameters distribution is not uniform. We find that biases do not take a lot of memory and show that biases can be left uncompressed to improve the overall quality without affecting the compression rate. We also propose to use an error-feedback mechanism during retraining, to preserve the compressed model as a stale gradient. We empirically show that NMT models based on Transformer or RNN architecture can be compressed up to 4-bit precision without any noticeable quality degradation. Models can be compressed up to binary precision, albeit with lower quality. RNN architecture seems to be more robust towards compression, compared to the Transformer.
We present META-MT, a meta-learning approach to adapt Neural Machine Translation (NMT) systems in a few-shot setting. META-MT provides a new approach to make NMT models easily adaptable to many target do- mains with the minimal amount of in-domain data. We frame the adaptation of NMT systems as a meta-learning problem, where we learn to adapt to new unseen domains based on simulated offline meta-training domain adaptation tasks. We evaluate the proposed meta-learning strategy on ten domains with general large scale NMT systems. We show that META-MT significantly outperforms classical domain adaptation when very few in- domain examples are available. Our experiments shows that META-MT can outperform classical fine-tuning by up to 2.5 BLEU points after seeing only 4, 000 translated words (300 parallel sentences).
The article is focused on automatic development and ranking of a large corpus for Russian paraphrase generation which proves to be the first corpus of such type in Russian computational linguistics. Existing manually annotated paraphrase datasets for Russian are limited to small-sized ParaPhraser corpus and ParaPlag which are suitable for a set of NLP tasks, such as paraphrase and plagiarism detection, sentence similarity and relatedness estimation, etc. Due to size restrictions, these datasets can hardly be applied in end-to-end text generation solutions. Meanwhile, paraphrase generation requires a large amount of training data. In our study we propose a solution to the problem: we collect, rank and evaluate a new publicly available headline paraphrase corpus (ParaPhraser Plus), and then perform text generation experiments with manual evaluation on automatically ranked corpora using the Universal Transformer architecture.
Cross-lingual text summarization aims at generating a document summary in one language given input in another language. It is a practically important but under-explored task, primarily due to the dearth of available data. Existing methods resort to machine translation to synthesize training data, but such pipeline approaches suffer from error propagation. In this work, we propose an end-to-end cross-lingual text summarization model. The model uses reinforcement learning to directly optimize a bilingual semantic similarity metric between the summaries generated in a target language and gold summaries in a source language. We also introduce techniques to pre-train the model leveraging monolingual summarization and machine translation objectives. Experimental results in both English–Chinese and English–German cross-lingual summarization settings demonstrate the effectiveness of our methods. In addition, we find that reinforcement learning models with bilingual semantic similarity as rewards generate more fluent sentences than strong baselines.
The answer-agnostic question generation is a significant and challenging task, which aims to automatically generate questions for a given sentence but without an answer. In this paper, we propose two new strategies to deal with this task: question type prediction and copy loss mechanism. The question type module is to predict the types of questions that should be asked, which allows our model to generate multiple types of questions for the same source sentence. The new copy loss enhances the original copy mechanism to make sure that every important word in the source sentence has been copied when generating questions. Our integrated model outperforms the state-of-the-art approach in answer-agnostic question generation, achieving a BLEU-4 score of 13.9 on SQuAD. Human evaluation further validates the high quality of our generated questions. We will make our code public available for further research.
We evaluate the performance of transformer encoders with various decoders for information organization through a new task: generation of section headings for Wikipedia articles. Our analysis shows that decoders containing attention mechanisms over the encoder output achieve high-scoring results by generating extractive text. In contrast, a decoder without attention better facilitates semantic encoding and can be used to generate section embeddings. We additionally introduce a new loss function, which further encourages the decoder to generate high-quality embeddings.
Recent works have shown that Neural Machine Translation (NMT) models achieve impressive performance, however, questions about understanding the behavior of these models remain unanswered. We investigate the unexpected volatility of NMT models where the input is semantically and syntactically correct. We discover that with trivial modifications of source sentences, we can identify cases where unexpected changes happen in the translation and in the worst case lead to mistranslations. This volatile behavior of translating extremely similar sentences in surprisingly different ways highlights the underlying generalization problem of current NMT models. We find that both RNN and Transformer models display volatile behavior in 26% and 19% of sentence variations, respectively.
We propose a method for natural language generation, choosing the most representative output rather than the most likely output. By viewing the language generation process from the voting theory perspective, we define representativeness using range voting and a similarity measure. The proposed method can be applied when generating from any probabilistic language model, including n-gram models and neural network models. We evaluate different similarity measures on an image captioning task and a machine translation task, and show that our method generates longer and more diverse sentences, providing a solution to the common problem of short outputs being preferred over longer and more informative ones. The generated sentences obtain higher BLEU scores, particularly when the beam size is large. We also perform a human evaluation on both tasks and find that the outputs generated using our method are rated higher.
We explore best practices for training small, memory efficient machine translation models with sequence-level knowledge distillation in the domain adaptation setting. While both domain adaptation and knowledge distillation are widely-used, their interaction remains little understood. Our large-scale empirical results in machine translation (on three language pairs with three domains each) suggest distilling twice for best performance: once using general-domain data and again using in-domain data with an adapted teacher.
In this paper, we introduce a system built for the Duolingo Simultaneous Translation And Paraphrase for Language Education (STAPLE) shared task at the 4th Workshop on Neural Generation and Translation (WNGT 2020). We participated in the English-to-Japanese track with a Transformer model pretrained on the JParaCrawl corpus and fine-tuned in two steps on the JESC corpus and then the (smaller) Duolingo training corpus. First, during training, we find it is essential to deliberately expose the model to higher-quality translations more often during training for optimal translation performance. For inference, encouraging a small amount of diversity with Diverse Beam Search to improve translation coverage yielded marginal improvement over regular Beam Search. Finally, using an auxiliary filtering model to filter out unlikely candidates from Beam Search improves performance further. We achieve a weighted F1 score of 27.56% on our own test set, outperforming the STAPLE AWS translations baseline score of 4.31%.
What is given below is a brief description of the two systems, called gFCONV and c-VAE, which we built in a response to the 2020 Duolingo Challenge. Both are neural models that aim at disrupting a sentence representation the encoder generates with an eye on increasing the diversity of sentences that emerge out of the process. Importantly, we decided not to turn to external sources for extra ammunition, curious to know how far we can go while confining ourselves to the data released by Duolingo. gFCONV works by taking over a pre-trained sequence model, and intercepting the output its encoder produces on its way to the decoder. c-VAE is a conditional variational auto-encoder, seeking the diversity by blurring the representation that the encoder derives. Experiments on a corpus constructed out of the public dataset from Duolingo, containing some 4 million pairs of sentences, found that gFCONV is a consistent winner over c-VAE though both suffered heavily from a low recall.
We introduce our TMU system that is submitted to The 4th Workshop on Neural Generation and Translation (WNGT2020) to English-to-Japanese (En→Ja) track on Simultaneous Translation And Paraphrase for Language Education (STAPLE) shared task. In most cases machine translation systems generate a single output from the input sentence, however, in order to assist language learners in their journey with better and more diverse feedback, it is helpful to create a machine translation system that is able to produce diverse translations of each input sentence. However, creating such systems would require complex modifications in a model to ensure the diversity of outputs. In this paper, we investigated if it is possible to create such systems in a simple way and whether it can produce desired diverse outputs. In particular, we combined the outputs from forward and backward neural translation models (NMT). Our system achieved third place in En→Ja track, despite adopting only a simple approach.
In this paper, we propose a transfer learning based simultaneous translation model by extending BART. We pre-trained BART with Korean Wikipedia and a Korean news dataset, and fine-tuned with an additional web-crawled parallel corpus and the 2020 Duolingo official training dataset. In our experiments on the 2020 Duolingo test dataset, our submission achieves 0.312 in weighted macro F1 score, and ranks second among the submitted En-Ko systems.
This paper describes the ADAPT Centre’s submission to STAPLE (Simultaneous Translation and Paraphrase for Language Education) 2020, a shared task of the 4th Workshop on Neural Generation and Translation (WNGT), for the English-to-Portuguese translation task. In this shared task, the participants were asked to produce high-coverage sets of plausible translations given English prompts (input source sentences). We present our English-to-Portuguese machine translation (MT) models that were built applying various strategies, e.g. data and sentence selection, monolingual MT for generating alternative translations, and combining multiple n-best translations. Our experiments show that adding the aforementioned techniques to the baseline yields an excellent performance in the English-to-Portuguese translation task.
We present our submission to the Simultaneous Translation And Paraphrase for Language Education (STAPLE) challenge. We used a standard Transformer model for translation, with a crosslingual classifier predicting correct translations on the output n-best list. To increase the diversity of the outputs, we used additional data to train the translation model, and we trained a paraphrasing model based on the Levenshtein Transformer architecture to generate further synonymous translations. The paraphrasing results were again filtered using our classifier. While the use of additional data and our classifier filter were able to improve results, the paraphrasing model produced too many invalid outputs to further improve the output quality. Our model without the paraphrasing component finished in the middle of the field for the shared task, improving over the best baseline by a margin of 10-22 % weighted F1 absolute.
This paper describes our submission to the 2020 Duolingo Shared Task on Simultaneous Translation And Paraphrase for Language Education (STAPLE). This task focuses on improving the ability of neural MT systems to generate diverse translations. Our submission explores various methods, including N-best translation, Monte Carlo dropout, Diverse Beam Search, Mixture of Experts, Ensembling, and Lexical Substitution. Our main submission is based on the integration of multiple translations from multiple methods using Consensus Voting. Experiments show that the proposed approach achieves a considerable degree of diversity without introducing noisy translations. Our final submission achieves a 0.5510 weighted F1 score on the blind test set for the English-Portuguese track.
We describe our submission to the 2020 Duolingo Shared Task on Simultaneous Translation And Paraphrase for Language Education (STAPLE). We view MT models at various training stages (i.e., checkpoints) as human learners at different levels. Hence, we employ an ensemble of multi-checkpoints from the same model to generate translation sequences with various levels of fluency. From each checkpoint, for our best model, we sample n-Best sequences (n=10) with a beam width =100. We achieve an 37.57 macro F1 with a 6 checkpoint model ensemble on the official shared task test data, outperforming a baseline Amazon translation system of 21.30 macro F1 and ultimately demonstrating the utility of our intuitive method.
This paper describes the University of Maryland’s submission to the Duolingo Shared Task on Simultaneous Translation And Paraphrase for Language Education (STAPLE). Unlike the standard machine translation task, STAPLE requires generating a set of outputs for a given input sequence, aiming to cover the space of translations produced by language learners. We adapt neural machine translation models to this requirement by (a) generating n-best translation hypotheses from a model fine-tuned on learner translations, oversampled to reflect the distribution of learner responses, and (b) filtering hypotheses using a feature-rich binary classifier that directly optimizes a close approximation of the official evaluation metric. Combination of systems that use these two strategies achieves F1 scores of 53.9% and 52.5% on Vietnamese and Portuguese, respectively ranking 2nd and 4th on the leaderboard.
This paper presents the Johns Hopkins University submission to the 2020 Duolingo Shared Task on Simultaneous Translation and Paraphrase for Language Education (STAPLE). We participated in all five language tasks, placing first in each. Our approach involved a language-agnostic pipeline of three components: (1) building strong machine translation systems on general-domain data, (2) fine-tuning on Duolingo-provided data, and (3) generating n-best lists which are then filtered with various score-based techniques. In addi- tion to the language-agnostic pipeline, we attempted a number of linguistically-motivated approaches, with, unfortunately, little success. We also find that improving BLEU performance of the beam-search generated translation does not necessarily improve on the task metric—weighted macro F1 of an n-best list.
This paper describes the third place submission to the shared task on simultaneous translation and paraphrasing for language education at the 4th workshop on Neural Generation and Translation (WNGT) for ACL 2020. The final system leverages pre-trained translation models and uses a Transformer architecture combined with an oversampling strategy to achieve a competitive performance. This system significantly outperforms the baseline on Hungarian (27% absolute improvement in Weighted Macro F1 score) and Portuguese (33% absolute improvement) languages.
This paper describes the submissions of the NiuTrans Team to the WNGT 2020 Efficiency Shared Task. We focus on the efficient implementation of deep Transformer models (Wang et al., 2019; Li et al., 2019) using NiuTensor, a flexible toolkit for NLP tasks. We explored the combination of deep encoder and shallow decoder in Transformer models via model compression and knowledge distillation. The neural machine translation decoding also benefits from FP16 inference, attention caching, dynamic batching, and batch pruning. Our systems achieve promising results in both translation quality and efficiency, e.g., our fastest system can translate more than 40,000 tokens per second with an RTX 2080 Ti while maintaining 42.9 BLEU on newstest2018.
This paper describes the OpenNMT submissions to the WNGT 2020 efficiency shared task. We explore training and acceleration of Transformer models with various sizes that are trained in a teacher-student setup. We also present a custom and optimized C++ inference engine that enables fast CPU and GPU decoding with few dependencies. By combining additional optimizations and parallelization techniques, we create small, efficient, and high-quality neural machine translation models.
We participated in all tracks of the Workshop on Neural Generation and Translation 2020 Efficiency Shared Task: single-core CPU, multi-core CPU, and GPU. At the model level, we use teacher-student training with a variety of student sizes, tie embeddings and sometimes layers, use the Simpler Simple Recurrent Unit, and introduce head pruning. On GPUs, we used 16-bit floating-point tensor cores. On CPUs, we customized 8-bit quantization and multiple processes with affinity for the multi-core setting. To reduce model size, we experimented with 4-bit log quantization but use floats at runtime. In the shared task, most of our submissions were Pareto optimal with respect the trade-off between time and quality.
Recent studies have shown that translation quality of NMT systems can be improved by providing document-level contextual information. In general sentence-based NMT models are extended to capture contextual information from large-scale document-level corpora which are difficult to acquire. Domain adaptation on the other hand promises adapting components of already developed systems by exploiting limited in-domain data. This paper presents FJWU’s system submission at WNGT, we specifically participated in Document level MT task for German-English translation. Our system is based on context-aware Transformer model developed on top of original NMT architecture by integrating contextual information using attention networks. Our experimental results show providing previous sentences as context significantly improves the BLEU score as compared to a strong NMT baseline. We also studied the impact of domain adaptation on document level translationand were able to improve results by adaptingthe systems according to the testing domain.
We present the task of Simultaneous Translation and Paraphrasing for Language Education (STAPLE). Given a prompt in one language, the goal is to generate a diverse set of correct translations that language learners are likely to produce. This is motivated by the need to create and maintain large, high-quality sets of acceptable translations for exercises in a language-learning application, and synthesizes work spanning machine translation, MT evaluation, automatic paraphrasing, and language education technology. We developed a novel corpus with unique properties for five languages (Hungarian, Japanese, Korean, Portuguese, and Vietnamese), and report on the results of a shared task challenge which attracted 20 teams to solve the task. In our meta-analysis, we focus on three aspects of the resulting systems: external training corpus selection, model architecture and training decisions, and decoding and filtering strategies. We find that strong systems start with a large amount of generic training data, and then fine-tune with in-domain data, sampled according to our provided learner response frequencies.
Knowledge-based question answering (KB_QA) has long focused on simple questions that can be answered from a single knowledge source, a manually curated or an automatically extracted KB. In this work, we look at answering complex questions which often require combining information from multiple sources. We present a novel KB-QA system, Multique, which can map a complex question to a complex query pattern using a sequence of simple queries each targeted at a specific KB. It finds simple queries using a neural-network based model capable of collective inference over textual relations in extracted KB and ontological relations in curated KB. Experiments show that our proposed system outperforms previous KB-QA systems on benchmark datasets, ComplexWebQuestions and WebQuestionsSP.
Data augmentation methods are commonly used in computer vision and speech. However, in domains dealing with textual data, such techniques are not that common. Most of the existing methods rely on rephrasing, i.e. new sentences are generated by changing a source sentence, preserving its meaning. We argue that in tasks with opposable classes (such as Positive and Negative in sentiment analysis), it might be beneficial to also invert the source sentence, reversing its meaning, to generate examples of the opposing class. Methods that use somewhat similar intuition exist in the space of adversarial learning, but are not always applicable to text classification (in our experiments, some of them were even detrimental to the resulting classifier accuracy). We propose and evaluate two reversal-based methods on an NLI task of recognising a type of a simple logical expression from its description in plain-text form. After gathering a dataset on MTurk, we show that a simple heuristic using a notion of negating the main verb has a potential not only on its own, but that it can also boost existing state-of-the-art rephrasing-based approaches.
In this paper, we detail novel strategies for interpolating personalized language models and methods to handle out-of-vocabulary (OOV) tokens to improve personalized language models. Using publicly available data from Reddit, we demonstrate improvements in offline metrics at the user level by interpolating a global LSTM-based authoring model with a user-personalized n-gram model. By optimizing this approach with a back-off to uniform OOV penalty and the interpolation coefficient, we observe that over 80% of users receive a lift in perplexity, with an average of 5.4% in perplexity lift per user. In doing this research we extend previous work in building NLIs and improve the robustness of metrics for downstream tasks.
Many users communicate with chatbots and AI assistants in order to help them with various tasks. A key component of the assistant is the ability to understand and answer a user’s natural language questions for question-answering (QA). Because data can be usually stored in a structured manner, an essential step involves turning a natural language question into its corresponding query language. However, in order to train most natural language-to-query-language state-of-the-art models, a large amount of training data is needed first. In most domains, this data is not available and collecting such datasets for various domains can be tedious and time-consuming. In this work, we propose a novel method for accelerating the training dataset collection for developing the natural language-to-query-language machine learning models. Our system allows one to generate conversational multi-term data, where multiple turns define a dialogue session, enabling one to better utilize chatbot interfaces. We train two current state-of-the-art NL-to-QL models, on both an SQL and SPARQL-based datasets in order to showcase the adaptability and efficacy of our created data.
Text normalization and sanitization are intrinsic components of Natural Language Inferences. In Information Retrieval or Dialogue Generation, normalization of user queries or utterances enhances linguistic understanding by translating non-canonical text to its canonical form, on which many state-of-the-art language models are trained. On the other hand, text sanitization removes sensitive information to guarantee user privacy and anonymity. Existing approaches to normalization and sanitization mainly rely on hand-crafted heuristics and syntactic features of individual tokens while disregarding the linguistic context. Moreover, such context-unaware solutions cannot dynamically determine whether out-of-vocab tokens are misspelt or are entity names. In this work, we formulate text normalization and sanitization as a multi-task text generation approach and propose a neural hybrid pointer-generator network based on multi-head attention. Its generator effectively captures linguistic context during normalization and sanitization while its pointer dynamically preserves the entities that are generally missing in the vocabulary. Experiments show that our generation approach outperforms both token-based text normalization and sanitization, while the hybrid pointer-generator improves the generator-only baseline in terms of BLEU4 score, and classical attentional pointer networks in terms of pointing accuracy.
One of the core components of voice assistants is the Natural Language Understanding (NLU) model. Its ability to accurately classify the user’s request (or “intent”) and recognize named entities in an utterance is pivotal to the success of these assistants. NLU models can be challenged in some languages by code-switching or morphological and orthographic variations. This work explores the possibility of improving the accuracy of NLU models for Indic languages via the use of alternate representations of input text for NLU, specifically ISO-15919 and IndicSOUNDEX, a custom SOUNDEX designed to work for Indic languages. We used a deep neural network based model to incorporate the information from alternate representations into the NLU model. We show that using alternate representations significantly improves the overall performance of NLU models when training data is limited.
We consider the task of generating dialogue responses from background knowledge comprising of domain specific resources. Specifically, given a conversation around a movie, the task is to generate the next response based on background knowledge about the movie such as the plot, review, Reddit comments etc. This requires capturing structural, sequential and semantic information from the conversation context and the background resources. We propose a new architecture that uses the ability of BERT to capture deep contextualized representations in conjunction with explicit structure and sequence information. More specifically, we use (i) Graph Convolutional Networks (GCNs) to capture structural information, (ii) LSTMs to capture sequential information and (iii) BERT for the deep contextualized representations that capture semantic information. We analyze the proposed architecture extensively. To this end, we propose a plug-and-play Semantics-Sequences-Structures (SSS) framework which allows us to effectively combine such linguistic information. Through a series of experiments we make some interesting observations. First, we observe that the popular adaptation of the GCN model for NLP tasks where structural information (GCNs) was added on top of sequential information (LSTMs) performs poorly on our task. This leads us to explore interesting ways of combining semantic and structural information to improve the performance. Second, we observe that while BERT already outperforms other deep contextualized representations such as ELMo, it still benefits from the additional structural information explicitly added using GCNs. This is a bit surprising given the recent claims that BERT already captures structural information. Lastly, the proposed SSS framework gives an improvement of 7.95% on BLUE score over the baseline.
Contextualized word embeddings provide better initialization for neural networks that deal with various natural language understanding (NLU) tasks including Question Answering (QA) and more recently, Question Generation(QG). Apart from providing meaningful word representations, pre-trained transformer models (Vaswani et al., 2017), such as BERT (Devlin et al., 2019) also provide self-attentions which encode syntactic information that can be probed for dependency parsing (Hewitt and Manning, 2019) and POStagging (Coenen et al., 2019). In this paper, we show that the information from selfattentions of BERT are useful for language modeling of questions conditioned on paragraph and answer phrases. To control the attention span, we use semi-diagonal mask and utilize a shared model for encoding and decoding, unlike sequence-to-sequence. We further employ copy-mechanism over self-attentions to acheive state-of-the-art results for Question Generation on SQuAD v1.1 (Rajpurkar et al., 2016).
Dialog State Tracking (DST) is a problem space in which the effective vocabulary is practically limitless. For example, the domain of possible movie titles or restaurant names is bound only by the limits of language. As such, DST systems often encounter out-of-vocabulary words at inference time that were never encountered during training. To combat this issue, we present a targeted data augmentation process, by which a practitioner observes the types of errors made on held-out evaluation data, and then modifies the training data with additional corpora to increase the vocabulary size at training time. Using this with a RoBERTa-based Transformer architecture, we achieve state-of-the-art results in comparison to systems that only mask trouble slots with special tokens. Additionally, we present a data-representation scheme for seamlessly retargeting DST architectures to new domains.
Building conversational systems in new domains and with added functionality requires resource-efficient models that work under low-data regimes (i.e., in few-shot setups). Motivated by these requirements, we introduce intent detection methods backed by pretrained dual sentence encoders such as USE and ConveRT. We demonstrate the usefulness and wide applicability of the proposed intent detectors, showing that: 1) they outperform intent detectors based on fine-tuning the full BERT-Large model or using BERT as a fixed black-box encoder on three diverse intent detection data sets; 2) the gains are especially pronounced in few-shot setups (i.e., with only 10 or 30 annotated examples per intent); 3) our intent detectors can be trained in a matter of minutes on a single CPU; and 4) they are stable across different hyperparameter settings. In hope of facilitating and democratizing research focused on intention detection, we release our code, as well as a new challenging single-domain intent detection dataset comprising 13,083 annotated examples over 77 intents.
Task-oriented dialog models typically leverage complex neural architectures and large-scale, pre-trained Transformers to achieve state-of-the-art performance on popular natural language understanding benchmarks. However, these models frequently have in excess of tens of millions of parameters, making them impossible to deploy on-device where resource-efficiency is a major concern. In this work, we show that a simple convolutional model compressed with structured pruning achieves largely comparable results to BERT on ATIS and Snips, with under 100K parameters. Moreover, we perform acceleration experiments on CPUs, where we observe our multi-task model predicts intents and slots nearly 63x faster than even DistilBERT.
Neural dialogue models, despite their successes, still suffer from lack of relevance, diversity, and in many cases coherence in their generated responses. On the other hand, transformer-based models such as GPT-2 have demonstrated an excellent ability to capture long-range structures in language modeling tasks. In this paper, we present DLGNet, a transformer-based model for dialogue modeling. We specifically examine the use of DLGNet for multi-turn dialogue response generation. In our experiments, we evaluate DLGNet on the open-domain Movie Triples dataset and the closed-domain Ubuntu Dialogue dataset. DLGNet models, although trained with only the maximum likelihood objective, achieve significant improvements over state-of-the-art multi-turn dialogue models. They also produce best performance to date on the two datasets based on several metrics, including BLEU, ROUGE, and distinct n-gram. Our analysis shows that the performance improvement is mostly due to the combination of (1) the long-range transformer architecture with (2) the injection of random informative paddings. Other contributing factors include the joint modeling of dialogue context and response, and the 100% tokenization coverage from the byte pair encoding (BPE).
Speech-based virtual assistants, such as Amazon Alexa, Google assistant, and Apple Siri, typically convert users’ audio signals to text data through automatic speech recognition (ASR) and feed the text to downstream dialog models for natural language understanding and response generation. The ASR output is error-prone; however, the downstream dialog models are often trained on error-free text data, making them sensitive to ASR errors during inference time. To bridge the gap and make dialog models more robust to ASR errors, we leverage an ASR error simulator to inject noise into the error-free text data, and subsequently train the dialog models with the augmented data. Compared to other approaches for handling ASR errors, such as using ASR lattice or end-to-end methods, our data augmentation approach does not require any modification to the ASR or downstream dialog models; our approach also does not introduce any additional latency during inference time. We perform extensive experiments on benchmark data and show that our approach improves the performance of downstream dialog models in the presence of ASR errors, and it is particularly effective in the low-resource situations where there are constraints on model size or the training data is scarce.
Dialogue response generation models that use template ranking rather than direct sequence generation allow model developers to limit generated responses to pre-approved messages. However, manually creating templates is time-consuming and requires domain expertise. To alleviate this problem, we explore automating the process of creating dialogue templates by using unsupervised methods to cluster historical utterances and selecting representative utterances from each cluster. Specifically, we propose an end-to-end model called Deep Sentence Encoder Clustering (DSEC) that uses an auto-encoder structure to jointly learn the utterance representation and construct template clusters. We compare this method to a random baseline that randomly assigns templates to clusters as well as a strong baseline that performs the sentence encoding and the utterance clustering sequentially. To evaluate the performance of the proposed method, we perform an automatic evaluation with two annotated customer service datasets to assess clustering effectiveness, and a human-in-the-loop experiment using a live customer service application to measure the acceptance rate of the generated templates. DSEC performs best in the automatic evaluation, beats both the sequential and random baselines on most metrics in the human-in-the-loop experiment, and shows promising results when compared to gold/manually created templates.