Elamur Research⌕
☰
home / papers / 2021 / Decontextualization: Making Sentences Stand-Alone

Decontextualization: Making Sentences Stand-Alone

Paper · 2021 · Eunsol Choi, Jennimaria Palomaki, Matthew Lamm et al.

Summary

NLP systems such as summarizers, QA models and retrievers often lift single sentences out of documents, yet these excerpts frequently rely on surrounding text for references and scope. The authors formalize sentence decontextualization: rewriting a sentence, given its preceding context, so it is understandable alone while keeping its truth-conditional meaning (explicature). Implicatures need not survive. They turn this into a crowdsourcing protocol on English Wikipedia. Annotators label a sentence FEASIBLE or INFEASIBLE and then make small edits: swapping pronouns, completing names, dropping discourse markers, and adding bridging or scoping phrases. Using this data, they build a coreference-based system and a sequence-to-sequence generator. The largest generator produces good rewrites in most cases but still trails humans in thoroughness and accuracy. Pilot studies show that users value decontextualized answers as concise yet informative, and that indexing decontextualized sentences makes open-domain QA retrieval much cheaper. The task is a reusable building block for pipelines handling text outside its original context.

Key points

  • Definition: A rewrite s′ is valid if it is interpretable in an empty context and has the same truth-conditional meaning as the original sentence in its context. Extra-linguistic context (time, place, audience) is assumed unchanged, so expressions like "thus far" stay as they are.
  • Edit frequencies: A manual analysis of 200 examples counted the main edit types:
    • PRONOUN/NP SWAP: 40.5%
    • NAME COMPLETION: 11.5%
    • DM REMOVAL: 3.5%
    • Other types: BRIDGING, GLOBAL SCOPING and ADDITION
  • Models and evaluation: One model builds on the coreference model of Joshi et al. (2020), the other on the generation model of Raffel et al. (2019a). Outputs are scored with automatic metrics derived from Xu et al. (2016) and with human evaluation.
  • Retrieval result (NQ-open): Sentence-level indexing beats paragraph and 100-word segmentation on cost-effectiveness. Decontextualized sentences approach the recall of 100-word windows at about 1/10th of the computational cost.
  • Resource: The dataset and evaluation script are released at github.com/google-research/language (language/decontext). The current scope is English encyclopedic text rewritten for an empty context.

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

🏢GoogleMountain View · United States353 works · 364,806 citations🎓Stanford UniversityStanford · United States310 works · 284,480 citations🎓Columbia UniversityNew York · United States149 works · 48,852 citations🎓The University of Texas at AustinAustin · United States72 works · 55,722 citations

📚 Shelves 4 Computer Science · 9,086 works Artificial Intelligence · 8,889 works Arts and Humanities · 9,296 works Language and Linguistics · 9,206 works

Cite

Cite as Eunsol Choi, Jennimaria Palomaki, Matthew Lamm, Tom Kwiatkowski, Dipanjan Das and Michael Collins. 2021. Decontextualization: Making Sentences Stand-Alone. Transactions of the Association for Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.1162/tacl_a_00377

@article{choi2021decontextualization,
  title = {{Decontextualization: Making Sentences Stand-Alone}},
  author = {Eunsol Choi, Jennimaria Palomaki, Matthew Lamm, Tom Kwiatkowski, Dipanjan Das and Michael Collins},
  journal = {Transactions of the Association for Computational Linguistics},
  year = {2021},
  doi = {10.1162/tacl\_a\_00377},
  url = {https://doi.org/10.1162/tacl_a_00377}
}

References

26 references, 16 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Lora Aroyo and Chris Welty. 2015. Truth is a lie: Crowd truth and the seven myths of human annotation. AI Magazine, 36:15–24. DOI: https://doi.org/10.1609/aimag.v36i1 .2564 ↗ our card
  2. H. Paul Grice. 1975. Logic and conversation. In Peter Cole and Jerry L. Morgan, editors, Speech Acts, volume 3 of Syntax and Semantics, pages 41–58. Academic Press, New York. find
  3. Joachim Bingel, Gustavo Paetzold, and Anders Søgaard. 2018. Lexi: A tool for adaptive, per-Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat, and Ming-Wei Chang. 2020. Realm: Retrieval-augmented language model pre-training. DOI: https://doi.org/10.3115/1598819.1598830 ↗
  4. K. Humphreys, R. Gaizauskas, and Saliha Azzam. 1997. Event coreference for information extraction. DOI: https://doi.org/10.3115/1598819.1598830 ↗ our card
  5. Gautier Izacard and Edouard Grave. 2020. Leveraging passage retrieval with generative models for open domain question answering. our card
  6. Mandar Joshi, Danqi Chen, Yinhan Liu, Daniel S. Weld, Luke Zettlemoyer, and Omer Levy. 2020. SpanBERT: Improving pre-training by representing and predicting spans. Transactions of the Association for Computational Linguistics, 8:64–77. DOI: https://doi.org/10.1162/tacl a 00300 ↗ our card
  7. V. Karpukhin, B. Oğuz, S. Min, L. Wu, S. Edunov, D. Chen, and W.-T. Yih. 2020. Dense passage retrieval for Open-Domain question answering. Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP). DOI: https://doi.org/10 .18653/v1/2020.emnlp-main.550 our card
  8. Tom Kwiatkowski, Jennimaria Palomaki, Olivia Redfield, Michael Collins, Ankur Parikh, Chris Alberti, Danielle Epstein, Illia Polosukhin, Matthew Kelcey, Jacob Devlin, Kenton Lee, Kristina N. Toutanova, Llion Jones, Ming-Wei Chang, Andrew Dai, Jakob Uszkoreit, Quoc Le, and Slav Petrov. 2019. Natural… our card
  9. Kenton Lee, Ming-Wei Chang, and Kristina Toutanova. 2019. Latent retrieval for weakly supervised open domain question answering. arXiv preprint 1906.00300. https://doi.org/10.1145/584792 .584854 ↗ our card
  10. Karthik Narasimhan, Adam Yala, and Regina Barzilay. 2016. Improving information extraction by acquiring external evidence with reinforcement learning. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP). DOI: https://doi.org/10.18653/v1/D16 -1261 ↗
  11. Jahna Otterbacher, Dragomir R. Radev, and Airong Luo. 2002. Revisions that improve cohesion in multi-document summaries: a preliminary study. In Proceedings of the Annual Meeting of the Association for Computation Linguistics (ACL). DOI: https://doi.org/10.3115/1118162.1118166 ↗
  12. Ankur P. Parikh, Xuezhi Wang, Sebastian Gehrmann, Manaal Faruqui, Bhuwan Dhingra, Diyi Yang, and Dipanjan Das. 2020. ToTTo: A controlled table-to-text generation dataset. Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), abs/2004.14373. DOI: https://doi.org/… our card
  13. Ellie Pavlick and Tom Kwiatkowski. 2019. Inherent disagreements in human textual inferences. Transactions of the Association for Computational Linguistics, 7:677–694. DOI: https://doi.org/10.1162/tacl a 00293 ↗ our card
  14. Emily Pitler, Annie Louis, and Ani Nenkova. 2010. Automatic evaluation of linguistic quality in multi-document summarization. In Proceedings of the Annual Meeting of the Association for Computation Linguistics (ACL). find
  15. Sameer Pradhan, Alessandro Moschitti, Nianwen Xue, Olga Uryupina, and Yuchen Zhang. 2012. ConLL-2012 shared task: Modeling multilingual unrestricted coreference in ontonotes. In EMNLP-CoNLL Shared Task. our card
  16. Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J. Liu. 2019a. Exploring the limits of transfer learning with a unified text-to-text transformer. arXiv preprint 1910.10683. our card
  17. Junyi Jessy Li, Bridget O’Daniel, Y. Wu, W. Zhao, and A. Nenkova. 2016. Improving the annotation of sentence specificity. In LREC. find
  18. X. Liu and W. Croft. 2002. Passage retrieval based on language models. In CIKM ’02. find
  19. DOI: Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J. Liu. 2019b. Exploring the limits of transfer learning with a unified text-to-text transformer. ArXiv, abs/1910.10683. our card
  20. Josef Steinberger, Massimo Poesio, Mijail A. Kabadjov, and Karel Jeek. 2007. Two uses of anaphora resolution in summarization. Information Processing and Management, 43(6):1663–1680. DOI: https://doi.org/10.1016/j.ipm.2007.01.010 — [doi:10.1016/j.ipm.2007.01.010](https://doi.org/10.1016/j.ipm.2007.01.010) ↗
  21. Ina Rösiger, Arndt Riester, and Jonas Kuhn. 2018. Bridging resolution: Task definition, corpus resources and rule-based experiments. In Proceedings of the International Conference on Computational Linguistics (COLING). our card
  22. Preusz Susanne, Birte Schmitz, Christa Hauenschild, and Carla Umbach. 1992. Anaphora resolution in machine translation. find
  23. Wei Xu, Chris Callison-Burch, and Courtney Napoles. 2015. Problems in current text simplification research: New data can help. Transactions of the Association for Computational Linguistics, 3:283–297. DOI: https://doi.org/10.1162/tacl a 00139 ↗ our card
  24. Piyush Sharma, Nan Ding, Sebastian Goodman, and Radu Soricut. 2018. Conceptual captions: A cleaned, hypernymed, image alt-text dataset for automatic image captioning. In Proceedings of the Annual Meeting of the Association for Computation Linguistics (ACL). DOI: https://doi.org/10.18653/v1/P18 -1238… ↗ our card
  25. Wei Xu, Courtney Napoles, Ellie Pavlick, Quanze Chen, and Chris Callison-Burch. 2016. Optimizing statistical machine translation for text simplification. Transactions of the Association for Computational Linguistics, 4:401–415. DOI: https://doi.org/10.1162/tacl a 00107 ↗ our card
  26. Dan Sperber and Deirdre Wilson. 1986. Relevance: Communication and Cognition. Harvard University Press, USA. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/choi-2021-decontextualization-making-sentences-stand-alone/