Elamur Research⌕
☰
home / papers / 2023 / Less is More: Mitigate Spurious Correlations for Open-Domain Dialogue Response Generation Models by Causal Discovery

Less is More: Mitigate Spurious Correlations for Open-Domain Dialogue Response Generation Models by Causal Discovery

Paper · 2023 · Tao Feng, Lizhen Qu and Gholamreza Haffari

Summary

Open-domain response generators such as BlenderBot and DialoGPT write fluent replies but often lose track of long conversation histories. As a result, they drift into generic or off-topic answers. The authors trace part of this weakness to spurious correlations: history utterances that are statistically linked to a reply without actually causing it.

To study this, they built CGDialog, a small corpus in which annotators marked the utterances or clauses in ESConv and MSC dialogues that directly cause a given response. Perturbation tests show that fine-tuned models react more strongly to changes in non-cause utterances than to changes in the true causes.

Drawing on constraint-based causal discovery, the authors recast cause detection as conditional independence testing. A classifier judges whether an utterance depends on the response once the immediately preceding turn is taken into account. Because labeled data is scarce, the classifier is trained with ConSTrain, a constrained self-training scheme. It then removes non-causes from histories before generator training and reranks candidate responses at inference.

Human and automatic evaluations show gains in relevance, informativeness and fluency. The approach is model-agnostic.

Key points

  • CGDialog corpus: 160 dialogues (80 from ESConv, 80 from MSC) with 922 history–response pairs, 6,108 utterances and 2,872 direct-cause utterances. Agreement is high: Cohen's Kappa of 0.8149 at utterance level and F1 of 0.8449 at clause level.
  • Recency matters: utterances closer to the response are more likely to be direct causes. The immediately preceding turn is annotated as a cause in over 80% (ESConv) and 95% (MSC) of cases.
  • ConSTrain results: cause-identification F1 is 0.70 on CGDialog-ESConv and 0.73 on CGDialog-MSC. This beats the heuristics "Always u_{t-1}" and "Always u_{t-2}, u_{t-1}" as well as the ablations INIT, FC and IST. On balanced test sets, accuracy reaches 0.83 and 0.86.
  • Ablation findings: two constraints drive ConSTrain's performance.
    • Removing the context constraint, which discards pseudo-labels far from the response, causes the largest loss in recall and F1.
    • The threshold constraint improves both precision and recall.
  • Availability: the dataset, models and code are released at github.com/WilliamsToTo/CGDIALOG (TACL vol. 11, 2023).

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

🎓Monash UniversityMelbourne · Australia60 works · 54,866 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.Dialogue systems · 624 works

Cite

Cite as Tao Feng, Lizhen Qu and Gholamreza Haffari. 2023. Less is More: Mitigate Spurious Correlations for Open-Domain Dialogue Response Generation Models by Causal Discovery. 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_00561

@article{feng2023less,
  title = {{Less is More: Mitigate Spurious Correlations for Open-Domain Dialogue Response Generation Models by Causal Discovery}},
  author = {Tao Feng, Lizhen Qu and Gholamreza Haffari},
  journal = {Transactions of the Association for Computational Linguistics},
  year = {2023},
  doi = {10.1162/tacl\_a\_00561},
  url = {https://doi.org/10.1162/tacl_a_00561}
}

References

63 references, 23 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Reina Akama, Sho Yokoi, Jun Suzuki, and Kentaro Inui. 2020. Filtering noisy dialogue corpora by connectivity and content relatedness. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 941–958, Online. Association for Computational Linguistics. h… find
  2. Martin Popel, Matt Post, Raphael Rubino, Carolina Scarton, Lucia Specia, Marco Turchi, Karin Verspoor, and Marcos Zampieri. 2016. Findings of the 2016 conference on machine translation. In Proceedings of the First Conference on Machine Translation: Volume 2, Shared Task Papers, pages 131–198, Berlin…
  3. Reinald Kim Amplayo and Mirella Lapata. 2021. Informative and controllable opinion summarization. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 2662–2672, Online. Association for Computational Linguistics. find
  4. Guendalina Caldarini, Sardar Jaf, and Kenneth McGarry. 2022. A literature survey of recent advances in chatbots. Information, 13(1). https://doi.org/10.3390/info13010041 ↗
  5. Satanjeev Banerjee and Alon Lavie. 2005. METEOR: An automatic metric for MT evaluation with improved correlation with human judgments. In Proceedings of the ACL Workshop on Intrinsic and Extrinsic Evaluation Measures for Machine Translation and/or Summarization, pages 65–72, Ann Arbor, Michigan. Ass…
  6. Chris Callison-Burch, Cameron Fordyce, Philipp Koehn, Christof Monz, and Josh Schroeder. 2007. (meta-) evaluation of machine translation. In Proceedings of the Second Workshop on Statistical Machine Translation, pages 136–158, Prague, Czech Republic. Association for Computational Linguistics. https:…
  7. Siqi Bao, Huang He, Fan Wang, Hua Wu, and Haifeng Wang. 2020. PLATO: Pre-trained dialogue generation model with discrete latent variable. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 85–96, Online. Association for Computational Linguistics.
  8. Richárd Csáky, Patrik Purgai, and Gábor Recski. 2019. Improving neural conversational models with entropy-based data filtering. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 5650–5669, Florence, Italy. Association for Computational Linguistics.… find
  9. Jason Baumgartner, Savvas Zannettou, Brian Keegan, Megan Squire, and Jeremy Blackburn. 2020. The pushshift Reddit dataset. In Proceedings of the international AAAI conference on web and social media, volume 14, pages 830–839. https://doi.org/10.1609/icwsm.v14i1.7347 ↗
  10. Emily Dinan, Stephen Roller, Kurt Shuster, Angela Fan, Michael Auli, and Jason Weston. 2019. Wizard of Wikipedia: Knowledge-powered conversational agents.
  11. Rotem Dror, Lotem Peled-Cohen, Segev Shlomov, and Roi Reichart. 2020. Statistical Significance Testing for Natural Language Processing. Synthesis Lectures on Human Language Technologies. Morgan & Claypool Publishers. https://doi.org/10.1007/978-3-031-02174-9 ↗
  12. Alexis Bellot and Mihaela van der Schaar. 2019. Conditional independence testing using generative adversarial networks. https://doi.org/10.48550/arXiv.1907.04068 — [doi:10.48550/arxiv.1907.04068](https://doi.org/10.48550/arxiv.1907.04068) ↗
  13. Anja Belz and Eric Kow. 2010. Comparing rating scales and preference judgements in language evaluation. In Proceedings of the 6th International Natural Language Generation Conference. Association for Computational Linguistics. find
  14. Gautier Izacard and Edouard Grave. 2021. Leveraging passage retrieval with generative models for open domain question answering. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 874–880, Online. Association for Computa…
  15. Ondřej Bojar, Rajen Chatterjee, Christian Federmann, Yvette Graham, Barry Haddow, Matthias Huck, Antonio Jimeno Yepes, Philipp Koehn, Varvara Logacheva, Christof Monz, Matteo Negri, Aurélie Névéol, Mariana Neves, Diederik P. Kingma and Jimmy Ba. 2015. Adam: A method for stochastic optimization.…
  16. Svetlana Kiritchenko and Saif Mohammad. 2017. Best-worst scaling more reliable than rating scales: A case study on sentiment intensity annotation. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 465–470, Vancouver, Canada. As…
  17. Chia-Wei Liu, Ryan Lowe, Iulian Serban, Mike Noseworthy, Laurent Charlin, and Joelle Pineau. 2016. How NOT to evaluate your dialogue system: An empirical study of unsupervised evaluation metrics for dialogue response generation. In Proceedings of the 2016 Conference on Empirical Methods in Natural L…
  18. Nyoungwoo Lee and Ho-Jin Choi. 2022. Toward robust response selection model for cross negative sampling condition. In 2022 IEEE International Conference on Big Data and Smart Computing (BigComp), pages 395–397. find
  19. Siyang Liu, Chujie Zheng, Orianna Demasi, Sahand Sabour, Yu Li, Zhou Yu, Yong Jiang, and Minlie Huang. 2021. Towards emotional support dialog systems. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Lan…
  20. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Proce…
  21. 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. https://doi.org/10.48550/arXiv.1907.11692 ↗
  22. Jiwei Li, Michel Galley, Chris Brockett, Jianfeng Gao, and Bill Dolan. 2016. A diversity-promoting objective function for neural conversation models. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pag…
  23. David Lopez-Paz and Maxime Oquab. 2017. Revisiting classifier two-sample tests. In International Conference on Learning Representations. find
  24. Yanran Li, Hui Su, Xiaoyu Shen, Wenjie Li, Ziqiang Cao, and Shuzi Niu. 2017. DailyDialog: A manually labelled multi-turn dialogue dataset. In Proceedings of the Eighth International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 986–995. Taipei, Taiwan. Asian Federati…
  25. Brady Neal. 2020. Introduction to Causal Inference from a Machine Learning Perspective. find
  26. Zekang Li, Jinchao Zhang, Zhengcong Fei, Yang Feng, and Jie Zhou. 2021. Conversations are not flat: Modeling the dynamic information flow across dialogue utterances. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference… find
  27. Jordan J. Louviere, Terry N. Flynn, and Anthony Alfred John Marley. 2015. Best-Worst Scaling: Theory, Methods and Applications. Cambridge University Press. https://doi.org/10.1017/CBO9781107337855 — [doi:10.1017/cbo9781107337855](https://doi.org/10.1017/cbo9781107337855) ↗
  28. Ana Rita Nogueira, João Gama, and Carlos Abreu Ferreira. 2021. Causal discovery in machine learning: Theories and applications. Journal of Dynamics and Games, 8(3):203–231. https://doi.org/10.3934/jdg.20210085 man text using divergence frontiers. Advances in Neural Information Processing Systems,… — [doi:10.3934/jdg.20210085](https://doi.org/10.3934/jdg.20210085) ↗
  29. Soujanya Poria, Navonil Majumder, Devamanyu Hazarika, Deepanway Ghosal, Rishabh Bhardwaj, Samson Yu Bai Jian, Pengfei Hong, Romila Ghosh, Abhinaba Roy, Niyati Chhaya, Alexander Gelbukh, and Rada Mihalcea. 2021. Recognizing emotion cause in conversations. Cognitive Computation, 13(5):1317–1332. https… find
  30. Jekaterina Novikova, Ondřej Dušek, and Verena Rieser. 2018. RankME: Reliable human ratings for natural language generation. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers), pa…
  31. Ratish Puduppully and Mirella Lapata. 2021. Data-to-text generation with macro planning. Transactions of the Association for Computational Linguistics, 9:510–527. https://doi.org/10.1162/tacl_a_00381 ↗
  32. Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. BLEU: A method for automatic evaluation of machine translation. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pages 311–318, Philadelphia, Pennsylvania, USA. Association for Computational…
  33. Lisong Qiu, Juntao Li, Wei Bi, Dongyan Zhao, and Rui Yan. 2019. Are training samples correlated? Learning to generate dialogue responses with multiple references. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3826–3835, Florence, Italy. Association… find
  34. Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Köpf, Edward Yang, Zach DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang, Junjie Bai, and Sou… find
  35. Yingqi Qu, Yuchen Ding, Jing Liu, Kai Liu, Ruiyang Ren, Wayne Xin Zhao, Daxiang Dong, Hua Wu, and Haifeng Wang. 2021. RocketQA: An optimized training approach to dense passage retrieval for open-domain question answering. In Proceedings of the 2021 Conference of the North American Chapter of the Ass…
  36. Judea Pearl. 2009. Causality, Cambridge University Press. find
  37. Pranav Rajpurkar, Jian Zhang, Konstantin Lopyrev, and Percy Liang. 2016. SQuAD: 100,000+ questions for machine comprehension of text. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 2383–2392, Austin, Texas. Association for Computational Linguistics.…
  38. Judea Pearl and Thomas Verma. 1991. A theory of inferred causation. In Proceedings of the Second International Conference on Principles of Knowledge Representation and Reasoning, KR’91, pages 441–452, San Francisco, CA, USA. Morgan Kaufmann Publishers Inc. find
  39. Krishna Pillutla, Swabha Swayamdipta, Rowan Zellers, John Thickstun, Sean Welleck, Yejin Choi, and Zaid Harchaoui. 2021. Mauve: Measuring the gap between neural text and hu-Hannah Rashkin, Eric Michael Smith, Margaret Li, and Y-Lan Boureau. 2019. Towards empathetic open-domain conversation models: A… find
  40. Peter Spirtes, Clark N. Glymour, Richard Scheines, and David Heckerman. 2000. Causation, Prediction, and Search, MIT Press. find
  41. Julius Steen and Katja Markert. 2021. How to evaluate a summarizer: Study design and statistical analysis for manual linguistic quality evaluation. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 1861–1875, Online. As… find
  42. Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Eric Michael Smith, Y-Lan Boureau, and Jason Weston. 2021. Recipes for building an open-domain chatbot. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational…
  43. Yu Sun, Eric Tzeng, Trevor Darrell, and Alexei A. Efros. 2019. Unsupervised domain adaptation through self-supervision. find
  44. Xiangru Tang, Alexander Fabbri, Haoran Li, Ziming Mao, Griffin Adams, Borui Wang, Asli Celikyilmaz, Yashar Mehdad, and Dragomir Radev. 2022. Investigating crowdsourcing protocols for evaluating the factual consistency of summaries. In Proceedings of the 2022 Conference of the North American Chapter… find
  45. Shiori Sagawa, Aditi Raghunathan, Pang Wei Koh, and Percy Liang. 2020. An investigation of why overparameterization exacerbates spurious correlations. In International Conference on Machine Learning, pages 8346–8356. PMLR. find
  46. Chinnadhurai Sankar, Sandeep Subramanian, Chris Pal, Sarath Chandar, and Yoshua Bengio. 2019. Do neural dialog systems use the conversation history effectively? An empirical study. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 32–37, Florence, Ital… find
  47. Haohan Wang, Zeyi Huang, and Eric Xing. 2021. Learning robust models by countering spurious correlations. find
  48. Taesun Whang, Dongyub Lee, Dongsuk Oh, Chanhee Lee, Kijong Han, Dong-hun Lee, and Saebyeok Lee. 2021. In Do response selection models really know what’s next? Utterance manipulation strategies for multi-turn response selection. volume 35, pages 14041–14049. https://doi.org/10.1609/aaai.v35i16.17653 — [doi:10.1609/aaai.v35i16.17653](https://doi.org/10.1609/aaai.v35i16.17653) ↗
  49. Rajat Sen, Karthikeyan Shanmugam, Himanshu Asnani, Arman Rahimzamani, and Sreeram Kannan. 2018. Mimic and classify: A meta-algorithm for conditional independence testing. https://doi.org/10.48550/arXiv.1806.09708 ↗
  50. Rajat Sen, Ananda Theertha Suresh, Karthikeyan Shanmugam, Alexandros G. Dimakis, and Sanjay Shakkottai. 2017. Model-powered conditional independence test. Advances in Neural Information Processing Systems, 30. find
  51. Herbert A. Simon. 1954. Spurious correlation: A causal interpretation. Journal of the American Statistical Association, 49(267):467–479. https://doi.org/10.1080/01621459 .1954.10483515 — [doi:10.1080/01621459](https://doi.org/10.1080/01621459) ↗
  52. Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pierric Cistac, Tim Rault, Remi Louf, Morgan Funtowicz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, Mariama Drame, Quentin Lhoe… — Transformers: State-of-the-Art Natural Language Processing
  53. Jianfeng Gao, Jingjing Liu, and Bill Dolan. 2020. DIALOGPT: Large-scale generative pre-training for conversational response generation. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pages 270–278, Online. Association for Computatio…
  54. Thomas Wolf, Victor Sanh, Julien Chaumond, and Clement Delangue. 2019. TransferTransfo: A transfer learning approach for neural network based conversational agents. https://doi.org/10.48550/arXiv.1901.08149 ↗
  55. Chujie Zheng, Yong Liu, Wei Chen, Yongcai Leng, and Minlie Huang. 2021. CoMAE: A multi-factor hierarchical framework for empathetic response generation. In Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 813–824, Online. Association for Computational Linguistics. ht… find
  56. Jing Xu, Arthur Szlam, and Jason Weston. 2022. Beyond goldfish memory: Long-term open-domain conversation. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5180–5197, Dublin, Ireland. Association for Computational Linguistics.… find
  57. Hanxun Zhong, Zhicheng Dou, Yutao Zhu, Hongjin Qian, and Ji-Rong Wen. 2022. Less is more: Learning to refine dialogue history for personalized dialogue generation. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Tec… find
  58. Xi Yang, Xing He, Hansi Zhang, Yinghan Ma, Jiang Bian, and Yonghui Wu. 2020. Measurement of semantic textual similarity in clinical texts: Comparison of transformer-based models. JMIR Medical Informatics, 8(11): e19735. https://doi.org/10.2196/19735, PubMed: 33226350 ↗
  59. Saizheng Zhang, Emily Dinan, Jack Urbanek, Arthur Szlam, Douwe Kiela, and Jason Weston. 2018. Personalizing dialogue agents: I have a dog, do you have pets too? In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2204–2213, Melbou…
  60. Hao Zhou, Minlie Huang, Tianyang Zhang, Xiaoyan Zhu, and Bing Liu. 2018. In Emotional chatting machine: Emotional conversation generation with internal and external memory. Proceedings of the AAAI Conference on Artificial Intelligence. https://doi.org/10 .1609/aaai.v32i1.11325 find
  61. Yang Zou, Zhiding Yu, Xiaofeng Liu, BVK Kumar, and Jinsong Wang. 2019. Confidence regularized self-training. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 5982–5991. https://doi.org/10.1109/ICCV.2019 .00608 ↗
  62. Tianyi Zhang*, Varsha Kishore*, Felix Wu*, Kilian Q. Weinberger, and Yoav Artzi. 2020. Bertscore: Evaluating text generation with bert. find
  63. Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/feng-2023-less-is-more-mitigate-spurious/