Elamur Research⌕
☰
home / papers / 2020 / On Faithfulness and Factuality in Abstractive Summarization

On Faithfulness and Factuality in Abstractive Summarization

Paper · 2020 · Joshua Maynez, Shashi Narayan, Bernd Bohnet et al.

Summary

The paper investigates the extent to which abstractive neural summarization systems hallucinate content that is not present in the input document or that contradicts the facts. The authors, from Google Research, carried out a large-scale human annotation of summaries for the extreme single-sentence summarization task XSum. They evaluated the outputs of systems based on RNNs (PtGen), CNNs (TConvS2S) and Transformers (TranS2S, GPT-Tuned, BertS2S), as well as human-written reference summaries. They propose a division of hallucinations into intrinsic and extrinsic. Intrinsic hallucinations distort information that is present in the document. Extrinsic hallucinations add information that is absent from the source. Separately, the authors checked whether the hallucinated content is at least factually correct. Hallucinations were found in the majority of summaries from all systems. Models with pretrained parameters performed best, both in terms of ROUGE and in human evaluations. Another important finding is that the degree of textual entailment between the document and the summary reflects faithfulness better than ROUGE or BERTScore. It is therefore suitable as a basis for automatic metrics, as well as for training and decoding criteria.

Key points

  • Intrinsic or extrinsic hallucinations occur in more than 70% of single-sentence summaries. Extrinsic ones predominate, and more than 90% of them are erroneous.
  • The BertS2S model, initialized with pretrained parameters, produces the most faithful summaries. Among its extrinsic hallucinations, it also has the largest share of factually correct ones.
  • NLI/entailment-based metrics correlate more strongly with summary faithfulness than ROUGE and BERTScore. OpenIE-based metrics proved unsuitable for XSum: by these metrics, the systems barely differ.
  • If a summary is selected from the candidates produced by four systems according to the entailment score, the share of faithful summaries increases by approximately 5 percentage points, but ROUGE drops slightly. The ENTAIL→FAITH variant was fine-tuned on the faithfulness annotations (5-fold cross-validation). It yields a smaller gain in faithfulness but preserves ROUGE better.
  • The human annotations have been released in the google-research-datasets/xsum_hallucination_annotations repository on GitHub.

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 States430 works · 424,363 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 1Research subjects and languages the work is about.Summarization · 232 works

💡 Concepts 1Key ideas and methods explained on our concept pages.Text summarization · 49 works

Cite

Cite as Joshua Maynez, Shashi Narayan, Bernd Bohnet and Ryan McDonald. 2020. On Faithfulness and Factuality in Abstractive Summarization. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2020.acl-main.173 · arXiv 2005.00661

@inproceedings{maynez2020faithfulness,
  title = {{On Faithfulness and Factuality in Abstractive Summarization}},
  author = {Joshua Maynez, Shashi Narayan, Bernd Bohnet and Ryan McDonald},
  booktitle = {Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics},
  year = {2020},
  doi = {10.18653/v1/2020.acl-main.173},
  eprint = {2005.00661},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.18653/v1/2020.acl-main.173}
}

References

58 references, 35 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Chris Alberti, Daniel Andor, Emily Pitler, Jacob Devlin, and Michael Collins. 2019. Synthetic QA corpora generation with roundtrip consistency. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 6168– 6173, Florence, Italy.
  2. Kristjan Arumae and Fei Liu. 2019. Guiding extractive summarization with question-answering rewards. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2566–2577, Minneapolis, Minnesota. find
  3. Teun A. van Dijk and Walter Kintsch. 1978. Cognitive psychology and discourse: Recalling and summarizing stories. In Wolfgang U. Dressler, editor, Current Trends in Textlinguistics, pages 61–80. find
  4. Li Dong, Nan Yang, Wenhui Wang, Furu Wei, Xiaodong Liu, Yu Wang, Jianfeng Gao, Ming Zhou, and Hsiao-Wuen Hon. 2019. Unified language model pre-training for natural language understanding and generation. In Advances in Neural Information Processing Systems 32, pages 13063–13075. Curran Associates, In…
  5. Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2015. Neural machine translation by jointly learning to align and translate. In 3rd International Conference on Learning Representations, San Diego, CA, USA.
  6. Tobias Falke, Leonardo F. R. Ribeiro, Prasetya Ajie Utama, Ido Dagan, and Iryna Gurevych. 2019. Ranking generated summaries by correctness: An interesting but challenging application for natural language inference. In Proceedings of the 57th Annual Meeting of the Association for Computational Lingui…
  7. Michele Banko, Michael J. Cafarella, Stephen Soderland, Matt Broadhead, and Oren Etzioni. 2007. Open information extraction from the web. In Proceedings of the 20th International Joint Conference on Artifical Intelligence, pages 2670–2676, Hyderabad, India.
  8. Jonas Gehring, Michael Auli, David Grangier, Denis Yarats, and Yann N. Dauphin. 2017. Convolutional sequence to sequence learning. In Proceedings of the 34th International Conference on Machine Learning, volume 70, pages 1243—-1252, Sydney, NSW, Australia.
  9. Ann L. Brown and Jeanne D. Day. 1983. Macrorules for summarizing texts: The development of expertise. Journal of Verbal Learning and Verbal Behaviour, 22(1):1–14. find
  10. Sebastian Gehrmann, Yuntian Deng, and Alexander Rush. 2018. Bottom-up abstractive summarization. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 4098–4109, Brussels, Belgium.
  11. Yen-Chun Chen and Mohit Bansal. 2018. Fast abstractive summarization with reinforce-selected sentence rewriting. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, pages 675–686, Melbourne, Australia.
  12. Zihang Dai, Zhilin Yang, Yiming Yang, Jaime Carbonell, Quoc Le, and Ruslan Salakhutdinov. 2019. Transformer-XL: Attentive language models beyond a fixed-length context. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 2978–2988, Florence, Italy. find
  13. Hoa Trang Dang. 2005. Overview of DUC 2005. In Proceedings of the Document Understanding Conference, pages 1–12. find
  14. 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,…
  15. Bhuwan Dhingra, Manaal Faruqui, Ankur Parikh, Ming-Wei Chang, Dipanjan Das, and William Cohen. 2019. Handling divergent reference texts when evaluating table-to-text generation. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4884–4895, Florence, Ita…
  16. Ben Goodrich, Vinay Rao, Peter J. Liu, and Mohammad Saleh. 2019. Assessing the factual accuracy of generated text. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 166–175, New York, NY, USA. find
  17. Jiatao Gu, Zhengdong Lu, Hang Li, and Victor O.K. Li. 2016. Incorporating copying mechanism in sequence-to-sequence learning. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, pages 1631–1640, Berlin, Germany.
  18. Hardy, Shashi Narayan, and Andreas Vlachos. 2019. HighRES: Highlight-based reference-less evaluation of summarization. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3381–3392, Florence, Italy. find
  19. Karl Moritz Hermann, Tomáš Kočiský, Edward Grefenstette, Lasse Espeholt, Will Kay, Mustafa Suleyman, and Phil Blunsom. 2015. Teaching machines to read and comprehend. In Advances in Neural Information Processing Systems 28, pages 1693–1701. Curran Associates, Inc.
  20. Ari Holtzman, Jan Buys, Maxwell Forbes, and Yejin Choi. 2020. The curious case of neural text degeneration. In Proceedings of the 8th International Conference on Learning Representations, Virtual Conference, Formerly Addis Ababa Ethiopia.
  21. Urvashi Khandelwal, Kevin Clark, Dan Jurafsky, and Lukasz Kaiser. 2019. Sample efficient text summarization using a single pre-trained transformer. CoRR, abs/1905.08836. find
  22. Wojciech Kryscinski, Nitish Shirish Keskar, Bryan Mc-Cann, Caiming Xiong, and Richard Socher. 2019a. Neural text summarization: A critical evaluation. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Langu…
  23. Wojciech Kryscinski, Bryan McCann, Caiming Xiong, and Richard Socher. 2019b. Evaluating the factual consistency of abstractive text summarization. CoRR, abs/1910.12840.
  24. Taku Kudo and John Richardson. 2018. Sentencepiece: A simple and language independent subword tokenizer and detokenizer for neural text processing. CoRR, abs/1808.06226.
  25. Tom Kwiatkowski, Jennimaria Palomaki, Olivia Red-field, Michael Collins, Ankur Parikh, Chris Alberti, Danielle Epstein, Illia Polosukhin, Jacob Devlin, Kenton Lee, Kristina Toutanova, Llion Jones, Matthew Kelcey, Ming-Wei Chang, Andrew M. Dai, Jakob Uszkoreit, Quoc Le, and Slav Petrov. 2019. Natural…
  26. J. Richard Landis and Gary G. Koch. 1977. The measurement of observer agreement for categorical data. Biometrics, 33(1):159–174. find
  27. Rémi Lebret, David Grangier, and Michael Auli. 2016. Neural text generation from structured data with application to the biography domain. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 1203–1213, Austin, Texas.
  28. Chin Yew Lin and Eduard Hovy. 2003. Automatic evaluation of summaries using n-gram co-occurrence statistics. In Proceedings of the 2003 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics, pages 150–157.
  29. Peter J. Liu, Mohammad Saleh, Etienne Pot, Ben Goodrich, Ryan Sepassi, Lukasz Kaiser, and Noam Shazeer. 2018. Generating wikipedia by summarizing long sequences. In Proceedings of the 6th International Conference on Learning Representations, Vancouver Canada.
  30. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. RoBERTa: A robustly optimized BERT pretraining approach. CoRR, abs/1907.11692.
  31. Inderjeet Mani. 2001. Automatic summarization, volume 3. John Benjamins Publishing. find
  32. Shashi Narayan, Shay B. Cohen, and Mirella Lapata. 2018a. Don’t give me the details, just the summary! Topic-aware convolutional neural networks for extreme summarization. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 1797–1807, Brussels, Belgium.
  33. Shashi Narayan, Shay B. Cohen, and Mirella Lapata. 2018b. Ranking sentences for extractive summarization with reinforcement learning. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1747–1759, Ne…
  34. Shashi Narayan, Gonçalo Simoes, Ji Ma, Hannah Craighead, and Ryan T. McDonald. 2020. QURI-OUS: Question generation pretraining for text generation. CoRR, abs/2004.11026. — [arXiv:2004.11026](https://arxiv.org/abs/2004.11026) find
  35. Ani Nenkova. 2005. Automatic Text Summarization of Newswire: Lessons Learned from the Document Understanding Conference. In Proceedings of the 20th National Conference on Artificial Intelligence -Volume 3, pages 1436–1441. find
  36. Ani Nenkova and Kathleen McKeown. 2011. Automatic summarization. Foundations and Trends in Information Retrieval, 5(2–3):103–233. find
  37. Ani Nenkova and Rebecca Passonneau. 2004. Evaluating content selection in summarization: The Pyramid method. In Proceedings of the Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics, pages 145–152, Boston, Massachusetts, USA. find
  38. Timothy Niven and Hung-Yu Kao. 2019. Probing neural network comprehension of natural language arguments. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4658–4664, Florence, Italy. find
  39. Ramakanth Pasunuru and Mohit Bansal. 2018. Multi-reward reinforced summarization with saliency and entailment. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 646–653, New Orleans, Louisiana.
  40. Romain Paulus, Caiming Xiong, and Richard Socher. 2018. A deep reinforced model for abstractive summarization. In Proceedings of the 6th International Conference on Learning Representations, Vancouver, BC, Canada.
  41. Alec Radford, Karthik Narasimhan, Tim Salimans, and Ilya Sutskever. 2018. Improving language understanding by generative pre-training. Technical report. find
  42. Alec Radford, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. Technical report. find
  43. Sascha Rothe, Shashi Narayan, and Aliaksei Severyn. 2020. Leveraging pre-trained checkpoints for sequence generation tasks. To appear in Transactions of the Association for Computational Linguistics, abs/1907.12461.
  44. Evan Sandhaus. 2008. The New York Times Annotated Corpus. Linguistic Data Consortium, Philadelphia, 6(12). find
  45. Natalie Schluter. 2017. The limits of automatic summarisation according to rouge. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics, pages 41– 45, Valencia, Spain.
  46. Abigail See, Peter J. Liu, and Christopher D. Manning. 2017. Get to the point: Summarization with pointer-generator networks. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 1073–1083, Vancouver, Canada. find
  47. Abigail See, Aneesh Pappu, Rohun Saxena, Akhila Yerukola, and Christopher D. Manning. 2019. Do massively pretrained language models make better storytellers? In Proceedings of the 23rd Conference on Computational Natural Language Learning, pages 843–861, Hong Kong, China. find
  48. Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, and Tie-Yan Liu. 2019. MASS: Masked sequence to sequence pre-training for language generation. In Proceedings of the 36th International Conference on Machine Learning, Long Beach, California.
  49. Ilya Sutskever, Oriol Vinyals, and Quoc V Le. 2014. Sequence to sequence learning with neural networks. In Advances in Neural Information Processing Systems 27, pages 3104–3112. Curran Associates, Inc.
  50. Sean Welleck, Jason Weston, Arthur Szlam, and Kyunghyun Cho. 2019. Dialogue natural language inference. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3731–3741, Florence, Italy.
  51. Adina Williams, Nikita Nangia, and Samuel Bowman. 2018. A broad-coverage challenge corpus for sentence understanding through inference. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1112–1122,…
  52. Sam Wiseman, Stuart Shieber, and Alexander Rush. 2017. Challenges in data-to-document generation. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 2253–2263, Copenhagen, Denmark.
  53. Yonghui Wu, Mike Schuster, Zhifeng Chen, Quoc V. Le, Mohammad Norouzi, Wolfgang Macherey, Maxim Krikun, Yuan Cao, Qin Gao, Klaus Macherey, Jeff Klingner, Apurva Shah, Melvin Johnson, Xiaobing Liu, Lukasz Kaiser, Stephan Gouws, Yoshikiyo Kato, Taku Kudo, Hideto Kazawa, Keith Stevens, George Kurian, N… find
  54. Zhilin Yang, Zihang Dai, Yiming Yang, Jaime G. Carbonell, Ruslan Salakhutdinov, and Quoc V. Le. 2019. XLNet: Generalized autoregressive pretraining for language understanding. CoRR, abs/1906.08237.
  55. Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q. Weinberger, and Yoav Artzi. 2020. BERTScore: Evaluating text generation with BERT. In Proceedings of the 8th International Conference on Learning Representations, Virtual Conference, Formerly Addis Ababa Ethiopia. find
  56. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Advances in Neural Information Processing Systems 30, pages 5998–6008. Curran Associates, Inc.
  57. Alex Wang, Kyunghyun Cho, and Michael Lewis. 2020. Asking and answering questions to evaluate the factual consistency of summaries. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, Virtual Conference, Formerly Seattle, USA. find
  58. Sean Welleck, Ilia Kulikov, Stephen Roller, Emily Dinan, Kyunghyun Cho, and Jason Weston. 2020. Neural text generation with unlikelihood training. In Proceedings of the 8th International Conference on Learning Representations, Virtual Conference, Formerly Addis Ababa Ethiopia.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/maynez-2020-faithfulness-factuality-abstractive-summarization/