Elamur Research⌕
☰
home / papers / 2023 / Selectively Answering Ambiguous Questions

Selectively Answering Ambiguous Questions

Paper · 2023 · Jeremy Cole, Michael Zhang, Daniel Gillick et al.

Summary

Trustworthy language models should decline to answer when unsure. Earlier work on confidence estimation mostly assumed the question itself is clear, so that only the model's knowledge could be lacking. This paper separates two sources of doubt. Epistemic uncertainty means not knowing the fact. Denotational uncertainty means not knowing what the asker intends, such as which year or which event. The authors propose a "disambiguate then answer" prompting scheme: PaLM first rewrites the question as an explicit interpretation, then either answers it or abstains. For selective QA they compare four confidence scores: the model's likelihood, self-verification, and two sampling-based measures. One sampling measure checks how often sampled answers repeat the greedy answer; the other checks how varied the samples are. The experiments use closed-book QA on Natural Questions, AmbigQA and SituatedQA, testing models of several sizes in both instruction-tuned and untuned versions. Counting agreement among samples gave the best-calibrated confidence, with the largest gains on ambiguous questions. This helps QA systems recognise when to abstain or ask for clarification.

Key points

  • Sampling scores use 10 outputs at temperature 0.5, compared by exact match after lowercasing and removing punctuation.
    • Repetition is the share of samples equal to the greedy (T=0) answer.
    • Diversity is 1 − (unique answers / number of samples).
  • Likelihood is somewhat informative about epistemic uncertainty but fails to detect denotational ambiguity.
  • Self-verification and offering an explicit "unknown" option were not effective for selective answering.
  • Sampling-based scores are especially helpful for instruction-tuned variants and for ambiguous questions from AmbigQA and SituatedQA.
  • Limitations: only the PaLM model family and a closed-book setting were tested, compute grows linearly with the number of samples, and prompt wording or longer explanations barely affected calibration.

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

🎓Duke UniversityDurham · United States18 works · 14,944 citations

📚 Shelves 4The 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 works

Cite

Cite as Jeremy Cole, Michael Zhang, Daniel Gillick, Julian Eisenschlos, Bhuwan Dhingra and Jacob Eisenstein. 2023. Selectively Answering Ambiguous Questions. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2023.emnlp-main.35

@inproceedings{cole2023selectively,
  title = {{Selectively Answering Ambiguous Questions}},
  author = {Jeremy Cole, Michael Zhang, Daniel Gillick, Julian Eisenschlos, Bhuwan Dhingra and Jacob Eisenstein},
  booktitle = {Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing},
  year = {2023},
  doi = {10.18653/v1/2023.emnlp-main.35},
  url = {https://doi.org/10.18653/v1/2023.emnlp-main.35}
}

References

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

  1. Sources of uncertainty in machine learning–a statisticians’ view. arXiv e-prints, pages arXiv–2305. find
  2. Ayush Agrawal, Lester Mackey, and Adam Tauman Kalai. 2023. Do language models know when they’re hallucinating references? arXiv preprint arXiv:2305.18248. find
  3. Aryaman Arora, Clara Meister, and Ryan Cotterell. 2022. Estimating the entropy of linguistic distributions. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 175–195, Dublin, Ireland. Association for Computational Linguistics. find
  4. Jannis Bulian, Christian Buck, Wojciech Gajewski, Benjamin Börschinger, and Tal Schuster. 2022. Tomayto, tomahto. beyond token-level answer equivalence for question answering evaluation. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 291–305, Abu Dha…
  5. Chi-Keung Chow. 1957. An optimum character recognition system using decision functions. IRE Transactions on Electronic Computers, EC-6(4):247–254. find
  6. Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, et al. 2022. Palm: Scaling language modeling with pathways. arXiv preprint arXiv:2204.02311.
  7. Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Eric Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, et al. 2022. Scaling instruction-finetuned language models. arXiv preprint arXiv:2210.11416.
  8. Shrey Desai and Greg Durrett. 2020. Calibration of pre-trained transformers. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 295–302, Online. Association for Computational Linguistics. find
  9. Li Dong, Chris Quirk, and Mirella Lapata. 2018. Confidence modeling for neural semantic parsing. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 743–753, Melbourne, Australia. Association for Computational Linguistics. find
  10. Ran El-Yaniv et al. 2010. On the foundations of noise-free selective classification. Journal of Machine Learning Research, 11(5). find
  11. Angela Fan, Mike Lewis, and Yann Dauphin. 2018. Hierarchical neural story generation. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 889–898, Melbourne, Australia. Association for Computational Linguistics. find
  12. Cornelia Gruber, Patrick Oliver Schenk, Malte Schierholz, Frauke Kreuter, and Göran Kauermann. 2023. find
  13. Chuan Guo, Geoff Pleiss, Yu Sun, and Kilian Q Weinberger. 2017. On calibration of modern neural networks. In International conference on machine learning, pages 1321–1330. PMLR. find
  14. Ari Holtzman, Jan Buys, Li Du, Maxwell Forbes, and Yejin Choi. 2020. The curious case of neural text degeneration. In International Conference on Learning Representations. find
  15. Abhyuday Jagannatha and Hong Yu. 2020. Calibrating structured output predictors for natural language processing. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 2078–2092, Online. Association for Computational Linguistics. find
  16. Zhengbao Jiang, Jun Araki, Haibo Ding, and Graham Neubig. 2021. How can we know when language models know? on the calibration of language models for question answering. Transactions of the Association for Computational Linguistics, 9:962–977. find
  17. Mandar Joshi, Eunsol Choi, Daniel Weld, and Luke Zettlemoyer. 2017. TriviaQA: A large scale distantly supervised challenge dataset for reading comprehension. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1601–1611, Vancouver…
  18. Saurav Kadavath, Tom Conerly, Amanda Askell, Tom Henighan, Dawn Drain, Ethan Perez, Nicholas Schiefer, Zac Hatfield Dodds, Nova DasSarma, Eli Tran-Johnson, et al. 2022. Language models (mostly) know what they know. arXiv preprint arXiv:2207.05221. find
  19. Amita Kamath, Robin Jia, and Percy Liang. 2020. Selective question answering under domain shift. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5684– 5696, Online. Association for Computational Linguistics. find
  20. Lorenz Kuhn, Yarin Gal, and Sebastian Farquhar. 2022. Clam: Selective clarification for ambiguous questions with large language models. arXiv preprint arXiv:2212.07769. find
  21. Lorenz Kuhn, Yarin Gal, and Sebastian Farquhar. 2023. Semantic uncertainty: Linguistic invariances for uncertainty estimation in natural language generation. arXiv preprint arXiv:2302.09664. find
  22. Volodymyr Kuleshov and Percy S Liang. 2015. Calibrated structured prediction. In Advances in Neural Information Processing Systems, volume 28. Curran Associates, Inc. find
  23. Aviral Kumar and Sunita Sarawagi. 2019. Calibration of encoder decoder models for neural machine translation. arXiv preprint arXiv:1903.00802. find
  24. Tom Kwiatkowski, Jennimaria Palomaki, Olivia Redfield, 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…
  25. Kenton Lee, Ming-Wei Chang, and Kristina Toutanova. 2019. Latent retrieval for weakly supervised open domain question answering. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 6086–6096, Florence, Italy. Association for Computational Linguistics.
  26. Stephanie Lin, Jacob Hilton, and Owain Evans. 2022. Teaching models to express their uncertainty in words. arXiv preprint arXiv:2205.14334. find
  27. Zhen Lin, Shubhendu Trivedi, and Jimeng Sun. 2023. Generating with confidence: Uncertainty quantification for black-box large language models. arXiv eprints, pages arXiv–2305. find
  28. Yao Lu, Max Bartolo, Alastair Moore, Sebastian Riedel, and Pontus Stenetorp. 2022. Fantastically ordered prompts and where to find them: Overcoming few-shot prompt order sensitivity. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), p…
  29. Potsawee Manakul, Adian Liusie, and Mark JF Gales. 2023. SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models. arXiv preprint arXiv:2303.08896. find
  30. Clara Meister, Afra Amini, Tim Vieira, and Ryan Cotterell. 2021. Conditional Poisson stochastic beams. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 664–681, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics. find
  31. Clara Meister, Tiago Pimentel, Gian Wiher, and Ryan Cotterell. 2023. Locally typical sampling. Transactions of the Association for Computational Linguistics, 11:102–121.
  32. Sewon Min, Julian Michael, Hannaneh Hajishirzi, and Luke Zettlemoyer. 2020. AmbigQA: Answering ambiguous open-domain questions. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 5783– 5797, Online. Association for Computational Linguistics. find
  33. Ian Osband, Zheng Wen, Seyed Mohammad Asghari, Vikranth Dwaracherla, Morteza Ibrahimi, Xiuyuan Lu, and Benjamin Van Roy. 2021. Epistemic neural networks. arXiv preprint arXiv:2107.08924. — [arXiv:2107.08924](https://arxiv.org/abs/2107.08924) find
  34. Jie Ren, Jiaming Luo, Yao Zhao, Kundan Krishna, Mohammad Saleh, Balaji Lakshminarayanan, and Peter J Liu. 2023. Out-of-distribution detection and selective generation for conditional language models. In The Eleventh International Conference on Learning Representations. find
  35. Ivan Stelmakh, Yi Luan, Bhuwan Dhingra, and Ming-Wei Chang. 2022. ASQA: Factoid questions meet long-form answers. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 8273–8288, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
  36. Neeraj Varshney and Chitta Baral. 2022. Model cascading: Towards jointly improving efficiency and accuracy of NLP systems. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 11007–11021, Abu Dhabi, United Arab Emirates. Association for Computational Ling… find
  37. Neeraj Varshney, Swaroop Mishra, and Chitta Baral. 2022. Investigating selective prediction approaches across several tasks in IID, OOD, and adversarial settings. In Findings of the Association for Computational Linguistics: ACL 2022, pages 1995–2002, Dublin, Ireland. Association for Computational L… find
  38. Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc Le, Ed Chi, and Denny Zhou. 2022. Self-consistency improves chain of thought reasoning in language models. arXiv preprint arXiv:2203.11171.
  39. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, brian ichter, Fei Xia, Ed H. Chi, Quoc V Le, and Denny Zhou. 2022. Chain of thought prompting elicits reasoning in large language models. In Advances in Neural Information Processing Systems.
  40. Allen R. Wilcox. 1973. Indices of qualitative variation and political measurement. The Western Political Quarterly, 26(2):325–343. find
  41. John M Zelle and Raymond J Mooney. 1996. Learning to parse database queries using inductive logic programming. In Proceedings of the National Conference on Artificial Intelligence, pages 1050–1055.
  42. Hugh Zhang, Daniel Duckworth, Daphne Ippolito, and Arvind Neelakantan. 2021a. Trading off diversity and quality in natural language generation. In Proceedings of the Workshop on Human Evaluation of NLP Systems (HumEval), pages 25–33, Online. Association for Computational Linguistics.
  43. Michael Zhang and Eunsol Choi. 2021. SituatedQA: Incorporating extra-linguistic contexts into QA. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 7371–7387, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics. precisio… find
  44. Kaitlyn Zhou, Dan Jurafsky, and Tatsunori Hashimoto. 2023. Navigating the grey area: Expressions of overconfidence and uncertainty in language models. arXiv preprint arXiv:2302.13439. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/cole-2023-selectively-answering-ambiguous-questions/