A Factored Neural Network Model for Characterizing Online Discussions in Vector Space

Hao Cheng, Hao Fang, Mari Ostendorf


Abstract
We develop a novel factored neural model that learns comment embeddings in an unsupervised way leveraging the structure of distributional context in online discussion forums. The model links different context with related language factors in the embedding space, providing a way to interpret the factored embeddings. Evaluated on a community endorsement prediction task using a large collection of topic-varying Reddit discussions, the factored embeddings consistently achieve improvement over other text representations. Qualitative analysis shows that the model captures community style and topic, as well as response trigger patterns.
Anthology ID:
D17-1243
Volume:
Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing
Month:
September
Year:
2017
Address:
Copenhagen, Denmark
Venue:
EMNLP
SIG:
SIGDAT
Publisher:
Association for Computational Linguistics
Note:
Pages:
2296–2306
Language:
URL:
https://www.aclweb.org/anthology/D17-1243
DOI:
10.18653/v1/D17-1243
Bib Export formats:
BibTeX MODS XML EndNote
PDF:
http://aclanthology.lst.uni-saarland.de/D17-1243.pdf