Elamur Research⌕
☰
home / papers / 2022 / End-to-End Neural Bridging Resolution

End-to-End Neural Bridging Resolution

Paper · 2022 · Hideo Kobayashi, Yufang Hou and Vincent Ng

Summary

Bridging resolution means finding anaphoric noun phrases tied to an earlier mention by a non-identity relation (for example, "ruling-party members" depends on "Japan") and linking each to that antecedent. The authors argue that the field's usual evaluations are unrealistic, because systems receive gold mentions or even gold anaphors. They also note that bridging models use weaker encoders than coreference models do. They take three leading resolvers and adapt each to work directly on raw documents: the rule system of Rösiger et al. (2018), the neural multi-task model of Yu and Poesio (2020), and the rules-plus-neural hybrid of Kobayashi and Ng (2021). They then swap the biLSTM encoders for BERT and SpanBERT. Testing on ISNotes and BASHI, they find that the end-to-end systems trail their gold-mention versions by only about 2–3 F-score points. Their strongest variants also reach state-of-the-art results when gold mentions are given. Perturbation experiments and a manual error analysis show what the best model relies on and where it fails. The work provides end-to-end baselines and argues that gold-mention evaluation should be dropped.

Key points

  • The corpora are ISNotes and BASHI, which hold just 50 OntoNotes documents each. OntoNotes itself has more than 2000 documents annotated for coreference.
  • The best system is H2(SBERT P), a hybrid score-based pipeline using a SpanBERT encoder. The LSTM-encoder version is more cautious: on ISNotes it outputs about half as many bridging pairs, with higher precision and lower recall.
  • Recall for recognising bridging anaphors stays low, at 31.6% on ISNotes and 27.5% on BASHI. Mention-detection failures explain only a small share of this (3% and 1.3% of gold anaphors). Anaphors with "the" are recovered least often (20% versus 25% for other anaphors on ISNotes).
  • On ISNotes, most precision errors are mentions wrongly labelled as bridging that are actually new (31%) or old/coreferent (21%). Another 16.7% of wrong pairs have the right anaphor but the wrong antecedent.
  • In the training-data perturbation tests, replacing nouns or adjectives/adverbs hurts the model more than replacing verbs.

Interesting ideas

  • I1. One span encoder and the lower feed-forward layers are shared between bridging and coreference scoring, and the weights of the two losses are tuned on dev data. Both tasks benefit from the same mention representations.
  • I2. Each mention can choose a dummy antecedent whose score is fixed at zero, so "no link" competes directly with real candidates and no separate threshold is needed.
  • I3. The hybrid pipeline applies precise hand-written rules first, each pairing an anaphor condition with an antecedent condition, and sends only the leftover mentions to the learned model.
  • I4. Perturbation analysis replaces a whole word class in training with either seen words or nonexistent words, then retrains and tests on clean data to see which word classes the model depends on.
  • I5. Errors are broken down by information status (new/old/mediated) and by determiner type (definite, other determiner, bare NP), instead of reporting one F-score.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (frames): our coreference module could score "same entity" and "related entity" links over one shared mention representation, so part-of and member-of links appear next to identity chains.
  • I2 (frames): coreference and bridging decisions could always include an explicit "none" option, which makes it easy to explain why a mention stayed unlinked.
  • I3 (world): rules backed by WordNet/ontology part-whole relations could resolve "the window" to "the house" explainably, leaving only the hard cases to a statistical scorer.
  • I4 (learn): this suits our negative controls, for example checking whether a parser or frame learner leans on lexical nouns rather than structure.
  • I5 (learn): slicing coreference or tree-parsing evaluation by determiner and discourse status would show which mention types our silver data handles poorly.

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

 Institutions

🎓The University of Texas at DallasRichardson · United States47 works · 5,363 citations

📚 Shelves 5The 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 worksComputer Vision and Pattern Recognition · 1,282 works

🧭 Topics 1Research subjects and languages the work is about.Coreference resolution · 550 works

Cite

Cite as Hideo Kobayashi, Yufang Hou and Vincent Ng. 2022. End-to-End Neural Bridging Resolution. Proceedings of the 29th International Conference on Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{kobayashi2022end,
  title = {{End-to-End Neural Bridging Resolution}},
  author = {Hideo Kobayashi, Yufang Hou and Vincent Ng},
  booktitle = {Proceedings of the 29th International Conference on Computational Linguistics},
  year = {2022},
  doi = {10.48448/dcpr-0w86},
  url = {https://doi.org/10.48448/dcpr-0w86}
}

References

41 references, 30 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Amit Bagga and Breck Baldwin. 1998. Algorithms for scoring coreference chains. In Proceedings of the Linguistic Coreference Workshop at the First International Conference on Language Resources and Evaluation (LREC’98), pages 563–566, Granada, Spain. European Language Resources Association (ELRA).
  2. Herbert H. Clark. 1975. Bridging. In Proceedings of the 1975 Workshop on Theoretical Issues in Natural Language Processing, TINLAP ’75, page 169–174, USA. Association for Computational Linguistics. find
  3. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 4171–4186, Minneapolis, Minnesota. Association for Computational Linguistics. ↗
  4. Sepp Hochreiter and Jürgen Schmidhuber. 1997. Long short-term memory. Neural Computation, 9(8):1735– 1780. ↗
  5. Matthew Honnibal and Ines Montani. 2017. spaCy 2: Natural language understanding with Bloom embeddings, convolutional neural networks and incremental parsing. To appear. find
  6. Yufang Hou. 2020a. Bridging anaphora resolution as question answering. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 1428–1438, Online. Association for Computational Linguistics.
  7. Yufang Hou. 2020b. Fine-grained information status classification using discourse context-aware BERT. In Proceedings of the 28th International Conference on Computational Linguistics, pages 6101–6112, Barcelona, Spain (Online). International Committee on Computational Linguistics.
  8. Yufang Hou. 2021. End-to-end neural information status classification. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 1377–1388, Punta Cana, Dominican Republic. Association for Computational Linguistics. ↗
  9. Yufang Hou, Katja Markert, and Michael Strube. 2013. Cascading collective classification for bridging anaphora recognition using a rich linguistic feature set. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 814–820, Seattle, Washington, USA. Association for Computational Linguistics.
  10. Yufang Hou, Katja Markert, and Michael Strube. 2014. A rule-based system for unrestricted bridging resolution: Recognizing bridging anaphora and finding links to antecedents. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 2082–2093, Doha, Qatar. Association for Computational Linguistics.
  11. Yufang Hou, Katja Markert, and Michael Strube. 2018. Unrestricted bridging resolution. Computational Linguistics, 44(2):237–284. ↗
  12. Mandar Joshi, Danqi Chen, Yinhan Liu, Daniel S. Weld, Luke Zettlemoyer, and Omer Levy. 2020. Span-BERT: Improving pre-training by representing and predicting spans. Transactions of the Association for Computational Linguistics, 8:64–77.
  13. Mandar Joshi, Omer Levy, Luke Zettlemoyer, and Daniel Weld. 2019. BERT for coreference resolution: Baselines and analysis. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 5803–5808, Hong Kong, China. Association for Computational Linguistics. ↗
  14. Sopan Khosla, Juntao Yu, Ramesh Manuvinakurike, Vincent Ng, Massimo Poesio, Michael Strube, and Carolyn Rosé. 2021. find
  15. The CODI-CRAC 2021 shared task on anaphora, bridging, and discourse deixis in dialogue. In Proceedings of the CODI-CRAC 2021 Shared Task on Anaphora, Bridging, and Discourse Deixis in Dialogue, pages 1–15, Punta Cana, Dominican Republic. Association for Computational Linguistics.
  16. Hongjin Kim, Damrin Kim, and Harksoo Kim. 2021. The pipeline model for resolution of anaphoric reference and resolution of entity reference. In Proceedings of the CODI-CRAC 2021 Shared Task on Anaphora, Bridging, and Discourse Deixis in Dialogue, pages 43–47, Punta Cana, Dominican Republic. Association for Computational Linguistics.
  17. Hideo Kobayashi, Yufang Hou, and Vincent Ng. 2022. Constrained multi-task learning for bridging resolution. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 759–770, Dublin, Ireland. Association for Computational Linguistics. find
  18. Hideo Kobayashi, Shengjie Li, and Vincent Ng. 2021. Neural anaphora resolution in dialogue. In Proceedings of the CODI-CRAC 2021 Shared Task on Anaphora, Bridging, and Discourse Deixis in Dialogue, pages 16–31, Punta Cana, Dominican Republic. Association for Computational Linguistics. ↗
  19. Hideo Kobayashi and Vincent Ng. 2020. Bridging resolution: A survey of the state of the art. In Proceedings of the 28th International Conference on Computational Linguistics, pages 3708–3721, Barcelona, Spain (Online). International Committee on Computational Linguistics. ↗
  20. Hideo Kobayashi and Vincent Ng. 2021. Bridging resolution: Making sense of the state of the art. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1652–1659, Online. Association for Computational Linguistics. ↗
  21. Shengjie Li, Hideo Kobayashi, and Vincent Ng. 2022. Neural anaphora resolution in dialogue revisited. In Proceedings of the CODI-CRAC 2022 Shared Task on Anaphora, Bridging, and Discourse Deixis in Dialogue, Gyeongju, Republic of Korea. Association for Computational Linguistics.
  22. Xiaoqiang Luo. 2005. On coreference resolution performance metrics. In Proceedings of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing, pages 25–32, Vancouver, British Columbia, Canada. Association for Computational Linguistics. ↗
  23. Christopher Manning, Mihai Surdeanu, John Bauer, Jenny Finkel, Steven Bethard, and David McClosky. 2014. The Stanford CoreNLP natural language processing toolkit. In Proceedings of 52nd Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pages 55–60, Baltimore, Maryland. Association for Computational Linguistics. ↗
  24. Katja Markert, Yufang Hou, and Michael Strube. 2012. Collective classification for fine-grained information status. In Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 795–804, Jeju Island, Korea. Association for Computational Linguistics.
  25. Shachar Mirkin, Ido Dagan, and Sebastian Padó. 2010. Assessing the role of discourse references in entailment inference. In Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics, pages 1209–1219, Uppsala, Sweden. Association for Computational Linguistics. find
  26. Malvina Nissim, Shipra Dingare, Jean Carletta, and Mark Steedman. 2004. An annotation scheme for information status in dialogue. In Proceedings of the Fourth International Conference on Language Resources and Evaluation (LREC’04), Lisbon, Portugal. European Language Resources Association (ELRA). ↗
  27. Onkar Pandit, Pascal Denis, and Liva Ralaivola. 2020. Integrating knowledge graph embeddings to improve mention representation for bridging anaphora resolution. In Proceedings of the Third Workshop on Computational Models of Reference, Anaphora and Coreference, pages 55–67, Barcelona, Spain (online). Association for Computational Linguistics. find
  28. Massimo Poesio, Rahul Mehta, Axel Maroudas, and Janet Hitzeman. 2004. Learning to resolve bridging references. In Proceedings of the 42nd Annual Meeting of the Association for Computational Linguistics (ACL-04), pages 143–150, Barcelona, Spain.
  29. Sameer Pradhan, Xiaoqiang Luo, Marta Recasens, Eduard Hovy, Vincent Ng, and Michael Strube. 2014. Scoring coreference partitions of predicted mentions: A reference implementation. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 30–35, Baltimore, Maryland. Association for Computational Linguistics. ↗
  30. Ellen F. Prince. 1981. Toward a taxonomy of given-new information. In P. Cole, editor, Syntax and semantics: Vol. 14. Radical Pragmatics, pages 223–255. Academic Press, New York. find
  31. Altaf Rahman and Vincent Ng. 2011. Learning the information status of noun phrases in spoken dialogues. In Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, pages 1069–1080, Edinburgh, Scotland, UK. Association for Computational Linguistics.
  32. Altaf Rahman and Vincent Ng. 2012. Learning the fine-grained information status of discourse entities. In Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics, pages 798–807, Avignon, France. Association for Computational Linguistics.
  33. Ina Rösiger. 2018. BASHI: A corpus of Wall Street Journal articles annotated with bridging links. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018), Miyazaki, Japan. European Language Resources Association (ELRA). ↗
  34. Ina Rösiger, Arndt Riester, and Jonas Kuhn. 2018. Bridging resolution: Task definition, corpus resources and rule-based experiments. In Proceedings of the 27th International Conference on Computational Linguistics, pages 3516–3528, Santa Fe, New Mexico, USA. Association for Computational Linguistics.
  35. Bo-Hsiang Tseng, Shruti Bhargava, Jiarui Lu, Joel Ruben Antony Moniz, Dhivya Piraviperumal, Lin Li, and Hong Yu. 2021. CREAD: Combined resolution of ellipses and anaphora in dialogues. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 3390–3406, Online. Association for Computational Linguistics. ↗
  36. Olga Uryupina, Ron Artstein, Antonella Bristot, Federica Cavicchio, Francesca Delogu, Kepa Joseba Rodriguez, and Massimo Poesio. 2020. Annotating a broad range of anaphoric phenomena, in a variety of genres: the ARRAU corpus. Natural Language Engineering, 26(1):95–128. ↗
  37. Marc Vilain, John Burger, John Aberdeen, Dennis Connolly, and Lynette Hirschman. 1995. A model-theoretic coreference scoring scheme. In Sixth Message Understanding Conference (MUC-6): Proceedings of a Conference Held in Columbia, Maryland, November 6-8, 1995. ↗
  38. Juntao Yu, Bernd Bohnet, and Massimo Poesio. 2020. Neural mention detection. In Proceedings of the 12th Language Resources and Evaluation Conference, pages 1–10, Marseille, France. European Language Resources Association. ↗
  39. Juntao Yu, Sopan Khosla, Ramesh Manuvinakurike, Lori Levin, Vincent Ng, Massimo Poesio, Michael Strube, and Carolyn Rosé. 2022. find
  40. The CODI-CRAC 2022 shared task on anaphora, bridging, and discourse deixis in dialogue. In Proceedings of the CODI-CRAC 2022 Shared Task on Anaphora, Bridging, and Discourse Deixis in Dialogue, Gyeongju, Republic of Korea. Association for Computational Linguistics.
  41. Juntao Yu and Massimo Poesio. 2020. Multitask learning-based neural bridging reference resolution. In Proceedings of the 28th International Conference on Computational Linguistics, pages 3534–3546, Barcelona, Spain (Online). International Committee on Computational Linguistics. ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/kobayashi-2022-end-end-neural-bridging-resolution/