Elamur Research⌕
☰
home / papers / 2016 / Identifying Causal Relations Using Parallel Wikipedia Articles

Identifying Causal Relations Using Parallel Wikipedia Articles

Paper · 2016 · Christopher Hidey and Kathy McKeown

Summary

The paper tackles automatic detection of causality in text, focusing on AltLexes: open-class phrases like "prompting" or "thanks to" that signal causal links. Unlike the small closed set of explicit connectives in the Penn Discourse Tree Bank (PDTB), these phrases vary widely in form, and labeled examples are scarce. The authors build a monolingual parallel corpus by pairing same-title articles from English and Simple Wikipedia and aligning paraphrased sentences with similarity measures (WikNet variants and doc2vec). Wherever a known causal connective lines up with a different phrase in the paired text, that phrase becomes a candidate AltLex. This yields training data through distant supervision. They then train a causal classifier that combines connective features, KL-divergence (KLD) features derived from the parallel data, and lexical-semantic features from FrameNet, WordNet and VerbNet. Self-training (bootstrapping) adds further data. The final system gains 11.05 F-measure points over a naive baseline. This shows that context and semantic knowledge, rather than the marker text alone, are needed to recognize causality expressed in varied ways.

Key points

  • Data: English and Simple Wikipedia dumps from September 11, 2015, converted with WikiExtractor and processed with Stanford CoreNLP (lemmatization, constituent and dependency parsing, NER). The PDTB contains 102 explicit connectives, 28 of them causal, which provide the seed set.
  • For paraphrase alignment, maximum F1 was 0.4850 for WikNet, 0.5981 for WikNet with a mismatch penalty (λ = 0.75), 0.6226 for Doc2Vec and 0.6263 for the combined score.
  • Bootstrapping improves on the purely supervised model by 7.14 absolute points, and the gain is statistically significant (binomial test, p < 7·10⁻⁶). Development-set performance drops after 2 rounds, so self-training stops there.
  • In the semantic feature ablation, FrameNet contributes the most (F 59.61), while WordNet (16.94) and VerbNet (6.92) are weak alone but complementary.
  • The AltLex text alone (CONN) performs poorly. KLD features give higher precision and lexical-semantic features give higher recall, which helps the model handle rare connectives.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 9,086 worksArtificial Intelligence · 8,889 worksArts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 worksBiochemistry, Genetics and Molecular Biology · 116 worksMolecular Biology · 114 works

Cite

Cite as Christopher Hidey and Kathy McKeown. 2016. Identifying Causal Relations Using Parallel Wikipedia Articles. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/p16-1135

@inproceedings{hidey2016identifying,
  title = {{Identifying Causal Relations Using Parallel Wikipedia Articles}},
  author = {Christopher Hidey and Kathy McKeown},
  booktitle = {Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2016},
  doi = {10.18653/v1/p16-1135},
  url = {https://doi.org/10.18653/v1/p16-1135}
}

References

26 references, 13 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Or Biran and Kathleen McKeown. 2013. Aggregated Word Pair Features for Implicit Discourse Relation Disambiguation. Proceedings of ACL.
  2. Sasha Blair-Goldensohn, Kathleen McKeown, and Owen Rambow. 2007. Building and Refining Rhetorical-Semantic Relation Models. Proceedings of NAACL-HLT. find
  3. Nathanael Chambers and Dan Jurafsky. 2008. Unsupervised Learning of Narrative Event Chains. Stanford University. Stanford, CA 94305. find
  4. Quang Xuan Do, Yee Seng Chan, and Dan Roth. 2011. Minimally Supervised Event Causality Identification. Transactions of ACL, 3:329344. find
  5. William Hwang, Hannaneh Hajishirzi, Mari Ostendorf, and Wei Wu. 2015. Aligning Sentences from Standard Wikipedia to Simple Wikipedia. Proceedings of NAACL-HLT. find
  6. Yangfeng Ji and Jacob Eisenstein. 2015. One Vector is Not Enough: Entity-Augmented Distributed Semantics for Discourse Relations. Proceedings of EMNLP.
  7. Karin Kipper, Hoa Trang Dan, and Martha Palmer. 2000. Class-Based Construction of a Verb Lexicon. American Association for Artifical Intelligence. find
  8. Majid Laali and Leila Kosseim. 2014. Inducing Discourse Connectives from Parallel Texts. Proceedings of COLING: Technical Papers.
  9. Quoc Le and Tomas Mikolov. 2014. Distributed Representations of Sentences and Documents. Proceedings of ICML.
  10. Ziheng Lin, Hwee Tou Ng, and Min-Yen Kan. 2012. A PDTB-Styled End-to-End Discourse Parser. Department of Computer Science, National University of Singapore.
  11. Christopher D. Manning, Mihai Surdeanu, John Bauer, Jenny Finkel, Steven J. Bethard, and David Mc-Closky. 2014. The Stanford CoreNLP Natural Language Processing Toolkit. Proceedings of ACL: System Demonstrations.
  12. Daniel Marcu and Abdessamad Echihabi 2001. An Unsupervised Approach to Recognizing Discourse Relations. Proceedings of ACL.
  13. George A. Miller. 1995. WordNet: A Lexical Database for English. Communications of the ACM.
  14. The PDTB Research Group 2008. The PDTB 2.0. Annotation Manual. Technical Report IRCS-08-01. Institute for Research in Cognitive Science, University of Pennsylvania. find
  15. Emily Pitler, Annie Louis, and Ani Nenkova. 2009. Automatic sense prediction for implicit discourse relations in text Proceedings of ACL.
  16. Emily Pitler and Ani Nenkova. 2009. Using Syntax to Disambiguate Explicit Discourse Connectives in Text. Proceedings of ACL-IJCNLP Short Papers.
  17. Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Miltsakaki, Livio Robaldo, Aravind Joshi and Bonnie Webber. 2008. The Penn Discourse Treebank 2.0. Proceedings of LREC.
  18. Rashmi Prasad, Aravind Joshi, and Bonnie Webber. 2010. Realization of Discourse Relations by Other Means: Alternative Lexicalizations. Proceedings of COLING.
  19. Kira Radinsky and Eric Horvitz. 2013. Mining the Web to Predict Future Events. Proceedings of WSDM. find
  20. Mehwish Riaz and Roxana Girju. 2013. Toward a Better Understanding of Causality between Verbal Events: Extraction and Analysis of the Causal Power of Verb-Verb Associations. Proceedings of SIGDIAL. find
  21. Mehwish Riaz and Roxana Girju. 2014. Recognizing Causality in Verb-Noun Pairs via Noun and Verb Semantics. Proceedings of the EACL 2014 Workshop on Computational Approaches to Causality in Language. find
  22. Josef Ruppenhofer, Michael Ellsworth, Miriam R. L. Petruck, Christopher R. Johnson, and Jan Scheffczyk. 2010. FrameNet II: Extended Theory and Practice. University of California, Berkeley. find
  23. Vladimir N. Vapnik. 1998. Statistical Learning Theory. Wiley-Interscience. find
  24. Juri Ganitkevitch, Benjamin Van Durme, and Chris Callison-Burch. 2013. PPDB: The Paraphrase Database. Proceedings of NAACL-HLT.
  25. F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, E. Duchesnay. 2011. Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research Volume 12. find
  26. Philipp Koehn, Hieu Hoang, Alexandra Birch, Chris Callison-Burch, Marcello Federico, Nicola Bertoldi, Brooke Cowan, Wade Shen, Christine Moran, Richard Zens, Chris Dyer, Ondej Bojar, Alexandra Constantin, and Evan Herbst. 2007. Moses: open source toolkit for statistical machine translation. Proceedi… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/hidey-2016-identifying-causal-relations-parallel-wikipedia/