Elamur Research⌕
☰
home / papers / 2022 / Generating Data to Mitigate Spurious Correlations in Natural Language Inference Datasets

Generating Data to Mitigate Spurious Correlations in Natural Language Inference Datasets

Paper · 2022 · Yuxiang Wu, Matt Gardner, Pontus Stenetorp et al.

Summary

NLI models often learn shortcuts: simple, task-irrelevant cues in the training data that happen to predict labels. These cues fail outside the training distribution. Instead of changing model architectures or loss functions, the authors keep the model fixed and replace its training data with a debiased version.

They fine-tune GPT-2 to produce premise, label and hypothesis in sequence. Unlikelihood training against perturbed labels improves label consistency, and a confidence filter from a trained NLI model removes weak samples. They then use the z-statistics framework of Gardner et al. (2021) to measure how strongly simple features correlate with labels. An iterative procedure called z-filtering discards instances that carry the currently most biased features. The same signal can also fine-tune the generator.

Debiased SNLI and MNLI variants improve generalisation on hard, out-of-distribution and adversarial test sets. Combined with product-of-experts, the method sets new best results on SNLI-hard and MNLI-hard. This shows that data-centric debiasing complements model-centric methods.

Key points

  • The bias features include unigrams/bigrams, hypothesis length, the hypothesis-to-premise length ratio, lexical overlap, and predictions of a BERT-base hypothesis-only model. The approach is not tied to this particular feature set.
  • Compared with training on the original data, gains reach 2.8pp on hypothesis-only bias sets, 13.3pp on syntactic bias (HANS), and 4.2pp on average across the adversarial suite of Liu et al. (2020b).
  • Conditional z-filtering starts from a seed dataset, so it can act as targeted augmentation. Example datasets include Seq-Z SNLI and Z-Aug MNLI.
  • The benefits carry over to larger models, with average gains of 2.30 (Roberta-base), 1.23 (Roberta-large) and 1.13 (Albert-xxlarge) points.
  • Code and generated datasets are available at github.com/jimmycode/gen-debiased-nli.

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

🔬Allen Institute for Artificial IntelligenceSeattle · United States219 works · 89,712 citations🎓University College LondonLondon · United Kingdom59 works · 16,463 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

🧭 Topics 1 Reasoning · 510 works

Cite

Cite as Yuxiang Wu, Matt Gardner, Pontus Stenetorp and Pradeep Dasigi. 2022. Generating Data to Mitigate Spurious Correlations in Natural Language Inference Datasets. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2022.acl-long.190

@inproceedings{wu2022generating,
  title = {{Generating Data to Mitigate Spurious Correlations in Natural Language Inference Datasets}},
  author = {Yuxiang Wu, Matt Gardner, Pontus Stenetorp and Pradeep Dasigi},
  booktitle = {Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2022},
  doi = {10.18653/v1/2022.acl-long.190},
  url = {https://doi.org/10.18653/v1/2022.acl-long.190}
}

References

43 references, 21 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Max Bartolo, Tristan Thrush, Robin Jia, Sebastian Riedel, Pontus Stenetorp, and Douwe Kiela. 2021. Improving question answering model robustness with synthetic adversarial data generation. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 8830– 8848, On… find
  2. Yonatan Belinkov, Adam Poliak, Stuart Shieber, Benjamin Van Durme, and Alexander Rush. 2019a. Don’t take the premise for granted: Mitigating artifacts in natural language inference. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 877–891, Florence, I… our card
  3. Yonatan Belinkov, Adam Poliak, Stuart Shieber, Benjamin Van Durme, and Alexander Rush. 2019b. On adversarial removal of hypothesis-only bias in natural language inference. In Proceedings of the Eighth Joint Conference on Lexical and Computational Semantics (*SEM 2019), pages 256–262, Minneapolis, Mi… our card
  4. Matt Gardner, William Merrill, Jesse Dodge, Matthew Peters, Alexis Ross, Sameer Singh, and Noah A. Smith. 2021. Competency problems: On finding and removing artifacts in language data. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 1801–1813, Online… find
  5. Prajjwal Bhargava, Aleksandr Drozd, and Anna Rogers. 2021. Generalization in NLI: Ways (not) to go beyond simple heuristics. In Proceedings of the Second Workshop on Insights from Negative Results in NLP, pages 125–135, Online and Punta Cana, Dominican Republic. Association for Computational Linguis… our card
  6. Mor Geva, Yoav Goldberg, and Jonathan Berant. 2019. Are we modeling the task or the annotator? an investigation of annotator bias in natural language understanding datasets. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conf… find
  7. Samuel R. Bowman. 2021. When combating hype, proceed with caution. CoRR, abs/2110.08300. find
  8. Samuel R. Bowman, Gabor Angeli, Christopher Potts, and Christopher D. Manning. 2015. A large annotated corpus for learning natural language inference. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 632–642, Lisbon, Portugal. Association for Computati… find
  9. Ronan Le Bras, Swabha Swayamdipta, Chandra Bhagavatula, Rowan Zellers, Matthew E. Peters, Ashish Sabharwal, and Yejin Choi. 2020. Adversarial filters of dataset biases. In Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13-18 July 2020, Virtual Event, volume 119 of P… find
  10. Danqi Chen, Jason Bolton, and Christopher D. Manning. 2016. A thorough examination of the CNN/Daily Mail reading comprehension task. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2358–2367, Berlin, Germany. Association for C… our card
  11. Christopher Clark, Mark Yatskar, and Luke Zettlemoyer. 2019. Don’t take the easy way out: Ensemble based methods for avoiding known dataset biases. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language… find
  12. 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,… our card
  13. Abbas Ghaddar, Phillippe Langlais, Mehdi Rezagholizadeh, and Ahmad Rashid. 2021. End-to-end self-debiasing framework for robust NLU training. In Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 1923–1929, Online. Association for Computational Linguistics. find
  14. Max Glockner, Vered Shwartz, and Yoav Goldberg. 2018. Breaking NLI systems with sentences that require simple lexical inferences. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 650–655, Melbourne, Australia. Association for… our card
  15. Suchin Gururangan, Swabha Swayamdipta, Omer Levy, Roy Schwartz, Samuel Bowman, and Noah A. Smith. 2018. Annotation artifacts in natural language inference data. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Techno… find
  16. He He, Sheng Zha, and Haohan Wang. 2019. Unlearn dataset bias in natural language inference by fitting the residual. In Proceedings of the 2nd Workshop on Deep Learning Approaches for Low-Resource NLP (DeepLo 2019), pages 132–142, Hong Kong, China. Association for Computational Linguistics. our card
  17. Geoffrey E Hinton. 2002. Training products of experts by minimizing contrastive divergence. Neural computation, 14(8):1771–1800. find
  18. Rabeeh Karimi Mahabadi, Yonatan Belinkov, and James Henderson. 2020. End-to-end bias mitigation by modelling biases in corpora. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 8706–8716, Online. Association for Computational Linguistics. our card
  19. Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, and Radu Soricut. 2020. ALBERT: A lite BERT for self-supervised on Computational Linguistics, pages 2340–2353, Santa Fe, New Mexico, USA. Association for Computational Linguistics. learning of language representations. In 8t… find
  20. Minwoo Lee, Seungpil Won, Juae Kim, Hwanhee Lee, Cheoneum Park, and Kyomin Jung. 2021. Crossaug: A contrastive data augmentation method for debiasing fact verification models. In Proceedings of the 30th ACM International Conference on Information & Knowledge Management, pages 3181–3185. find
  21. Patrick Lewis, Yuxiang Wu, Linqing Liu, Pasquale Minervini, Heinrich Küttler, Aleksandra Piktus, Pontus Stenetorp, and Sebastian Riedel. 2021. PAQ: 65 million probably-asked questions and what you can do with them. Transactions of the Association for Computational Linguistics, 9:1098–1115. our card
  22. Haochen Liu, Joseph Thekinen, Sinem Mollaoglu, Da Tang, Ji Yang, Youlong Cheng, Hui Liu, and Jiliang Tang. 2021. Toward annotator group bias in crowdsourcing. ArXiv preprint, abs/2110.08038. find
  23. Tianyu Liu, Zheng Xin, Baobao Chang, and Zhifang Sui. 2020a. HypoNLI: Exploring the artificial patterns of hypothesis-only bias in natural language inference. In Proceedings of the 12th Language Resources and Evaluation Conference, pages 6852– 6860, Marseille, France. European Language Resources Ass… our card
  24. Tianyu Liu, Zheng Xin, Xiaoan Ding, Baobao Chang, and Zhifang Sui. 2020b. An empirical study on model-agnostic debiasing strategies for robust natural language inference. In Proceedings of the 24th Conference on Computational Natural Language Learning, pages 596–608, Online. Association for Computat… our card
  25. 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. find
  26. Tom McCoy, Ellie Pavlick, and Tal Linzen. 2019. Right for the wrong reasons: Diagnosing syntactic heuristics in natural language inference. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3428–3448, Florence, Italy. Association for Computational Ling… our card
  27. Pasquale Minervini and Sebastian Riedel. 2018. Adversarially regularising neural NLI models to integrate logical background knowledge. In Proceedings of the 22nd Conference on Computational Natural Language Learning, pages 65–74, Brussels, Belgium. Association for Computational Linguistics. find
  28. Aakanksha Naik, Abhilasha Ravichander, Norman Sadeh, Carolyn Rose, and Graham Neubig. 2018. Stress test evaluation for natural language inference. In Proceedings of the 27th International Conference Yixin Nie, Yicheng Wang, and Mohit Bansal. 2019. Analyzing compositionality-sensitivity of NLI models… find
  29. Adam Poliak, Jason Naradowsky, Aparajita Haldar, Rachel Rudinger, and Benjamin Van Durme. 2018. Hypothesis only baselines in natural language inference. In Proceedings of the Seventh Joint Conference on Lexical and Computational Semantics, pages 180–191, New Orleans, Louisiana. Association for Compu… find
  30. Alexis Ross, Tongshuang Wu, Hao Peng, Matthew E Peters, and Matt Gardner. 2021. Tailor: Generating and perturbing text with semantic controls. ArXiv preprint, abs/2107.07150. our card
  31. Victor Sanh, Thomas Wolf, Yonatan Belinkov, and Alexander M. Rush. 2021. Learning from others’ mistakes: Avoiding dataset biases without modeling them. In 9th International Conference on Learning Representations, ICLR 2021, Virtual Event, Austria, May 3-7, 2021. OpenReview.net. find
  32. Timo Schick and Hinrich Schütze. 2021. Generating datasets with pretrained language models. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 6943– 6951, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics. our card
  33. Tal Schuster, Darsh Shah, Yun Jie Serene Yeo, Daniel Roberto Filizzola Ortiz, Enrico Santus, and Regina Barzilay. 2019. Towards debiasing fact verification models. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on… our card
  34. Roy Schwartz, Maarten Sap, Ioannis Konstas, Leila Zilles, Yejin Choi, and Noah A. Smith. 2017. The effect of different writing tasks on linguistic style: A case study of the ROC story cloze task. In Proceedings of the 21st Conference on Computational Natural Language Learning (CoNLL 2017), pages 15–… find
  35. Joe Stacey, Yonatan Belinkov, and Marek Rei. 2021. Supervising model attention with human explanations for robust natural language inference. ArXiv preprint, abs/2104.08142. — [arXiv:2104.08142](https://arxiv.org/abs/2104.08142) find
  36. Joe Stacey, Pasquale Minervini, Haim Dubossarsky, Sebastian Riedel, and Tim Rocktäschel. 2020. Avoiding the Hypothesis-Only Bias in Natural Language Inference via Ensemble Adversarial Training. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages… our card
  37. James Thorne, Andreas Vlachos, Christos Christodoulopoulos, and Arpit Mittal. 2018. FEVER: a large-scale dataset for fact extraction and VERification. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Vo… our card
  38. Prasetya Ajie Utama, Nafise Sadat Moosavi, and Iryna Gurevych. 2020. Mind the trade-off: Debiasing NLU models without degrading the in-distribution performance. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 8717–8729, Online. Association for Comput… our card
  39. Haohan Wang, Da Sun, and Eric P. Xing. 2019. What if we simply swap the two text fragments? A straightforward yet effective way to test the robustness of methods to confounding signals in nature language inference tasks. In The Thirty-Third AAAI Conference on Artificial Intelligence, AAAI 2019, Hono… find
  40. Sean Welleck, Ilia Kulikov, Stephen Roller, Emily Dinan, Kyunghyun Cho, and Jason Weston. 2020. Neural text generation with unlikelihood training. In 8th International Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020. OpenReview.net. our card
  41. 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, Volume 1 (Long Pa… find
  42. Yiben Yang, Chaitanya Malaviya, Jared Fernandez, Swabha Swayamdipta, Ronan Le Bras, Ji-Ping Wang, Chandra Bhagavatula, Yejin Choi, and Doug Downey. 2020. G-daug: Generative data augmentation for commonsense reasoning. In EMNLP (Findings), volume EMNLP 2020 of Findings of ACL, pages 1008–1025. Associ… our card
  43. Xiang Zhou and Mohit Bansal. 2020. Towards robustifying NLI models against lexical dataset biases. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 8759– 8771, Online. Association for Computational Linguistics. our card

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/wu-2022-generating-data-mitigate-spurious-correlations/