Elamur Research
☰
home / papers / 2020 / Towards Debiasing Sentence Representations

Towards Debiasing Sentence Representations

Written for the Research library (claude-opus-5-5) from the paper text. Original — aclanthology.org/2020.acl-main.488.pdf

Summary

Contextual sentence encoders such as BERT and ELMo absorb social stereotypes from their training text. Earlier debiasing methods targeted static word embeddings, and retraining huge sentence encoders each time a new bias is found is impractical. The authors propose SENT-DEBIAS, a post-hoc method that extends Hard-Debias (Bolukbasi et al., 2016) from words to sentences.

They begin with lists of attribute words: gender pairs and three-way religion tuples (Jewish/Christian/Muslim). These words are placed into naturally occurring sentences taken from several corpora covering written and spoken, formal and informal text. Swapping the attribute term produces parallel sentence sets. These sentences are encoded and centered, and PCA extracts a bias subspace. Any new representation is then cleaned by subtracting its projection onto that subspace.

On BERT and ELMo, the method lowers measured bias. Gains are larger when the templates are numerous and varied rather than short hand-written patterns. Downstream accuracy changes only slightly. The work matters because it offers an inexpensive, retraining-free route to fairer widely used sentence representations.

Key points

  • SENT-DEBIAS has four stages: define attribute words, contextualize them into sentences, estimate the bias subspace with PCA, and remove each representation's projection onto it.
  • Sentence templates come from WikiText-2, SST, Reddit, MELD and POM. Together they span encyclopedic text, movie reviews, forum posts, TV-comedy dialogue and spoken opinion videos.
  • Bias is assessed with association-style metrics (WEAT/SEAT, MAC). The authors caution that these metrics can reveal bias but cannot prove its absence.
  • Downstream results before and after debiasing:
    • BERT: SST-2 92.7→89.1, CoLA 57.6→55.4, QNLI 91.3→90.6.
    • ELMo: SST-2 89.6→90.0, CoLA 39.1→37.1.
  • Stated limitations:
    • Debiasing is applied to all sentences, including non-neutral ones.
    • The bias subspace must be re-estimated whenever BERT is fine-tuned.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Natural language processing

Cited by

112 papers of our library cite it (of the 140673 whose bibliographies we read); 8 of them have a page here, 104 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2020.acl-main.488

@inproceedings{liang2020debiasing,
  title = {{Towards Debiasing Sentence Representations}},
  author = {Liang, Paul Pu and Li, Irene Mengze and Zheng, Emily and Lim, Yao Chong and Salakhutdinov, Ruslan and Morency, Louis-Philippe},
  booktitle = {Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics},
  year = {2020},
  doi = {10.18653/v1/2020.acl-main.488},
  url = {https://doi.org/10.18653/v1/2020.acl-main.488}
}

References

51 references, 20 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Hervé Abdi and Lynne J. Williams. 2010. Principal component analysis. WIREs Comput. Stat., 2(4):433–459. find
  2. Emily Alsentzer, John Murphy, William Boag, Wei-Hung Weng, Di Jindi, Tristan Naumann, and Matthew McDermott. 2019. Publicly available clinical BERT embeddings. In Proceedings of the 2nd Clinical Natural Language Processing Workshop, pages 72–78, Minneapolis, Minnesota, USA. Association for Computati… find our card
  3. David Bamman, A. Seza Doğruöz, Jacob Eisenstein, Dirk Hovy, David Jurgens, Brendan O’Connor, Alice Oh, Oren Tsur, and Svitlana Volkova. 2016. Proceedings of the first workshop on NLP and computational social science. find
  4. Solon Barocas and Andrew D Selbst. 2016. Big data’s disparate impact. Calif. L. Rev., 104:671. find
  5. Christine Basta, Marta R. Costa-jussà, and Noe Casas. 2019. Evaluating the underlying gender bias in contextualized word embeddings. In Proceedings of the First Workshop on Gender Bias in Natural Language Processing, pages 33–39, Florence, Italy. Association for Computational Linguistics. find
  6. Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. 2016. Enriching word vectors with subword information. arXiv preprint arXiv:1607.04606. find
  7. Tolga Bolukbasi, Kai-Wei Chang, James Y Zou, Venkatesh Saligrama, and Adam T Kalai. 2016. Man is to computer programmer as woman is to homemaker? Debiasing word embeddings. In Proc. of NIPS, pages 4349–4357. find our card
  8. Shikha Bordia and Samuel R. Bowman. 2019. Identifying and reducing gender bias in word-level language models. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Student Research Workshop, pages 7–15, Minneapolis, Minnesota. Associati… find our card
  9. Aylin Caliskan, Joanna J Bryson, and Arvind Narayanan. 2017. Semantics derived automatically from language corpora contain human-like biases. Science. find
  10. Robert Dale. 2019. Law and word order: Nlp in legal tech. Natural Language Engineering, 25(1):211– 217. find
  11. 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,… find
  12. Sahaj Garg, Vincent Perot, Nicole Limtiaco, Ankur Taly, Ed H. Chi, and Alex Beutel. 2019. Counterfactual fairness in text classification through robustness. In Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society, AIES ’19, page 219–226, New York, NY, USA. Association for Computing… find
  13. Hila Gonen and Yoav Goldberg. 2019. Lipstick on a pig: Debiasing methods cover up systematic gender biases in word embeddings but do not remove them. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Vol… find our card
  14. Anthony Greenwald, T Poehlman, Eric Luis Uhlmann, and Mahzarin Banaji. 2009. Understanding and using the implicit association test: Iii. meta-analysis of predictive validity. find
  15. Kexin Huang, Jaan Altosaar, and Rajesh Ranganath. 2019. Clinicalbert: Modeling clinical notes and predicting hospital readmission. CoRR, abs/1904.05342. find our card
  16. Svetlana Kiritchenko and Saif Mohammad. 2018. Examining gender and race bias in two hundred sentiment analysis systems. In Proceedings of the Seventh Joint Conference on Lexical and Computational Semantics, pages 43–53, New Orleans, Louisiana. Association for Computational Linguistics. find our card
  17. Keita Kurita, Nidhi Vyas, Ayush Pareek, Alan W Black, and Yulia Tsvetkov. 2019. Measuring bias in contextualized word representations. In Proceedings of the First Workshop on Gender Bias in Natural Language Processing, pages 166–172, Florence, Italy. Association for Computational Linguistics. find our card
  18. Brian Larson. 2017. Gender as a variable in natural-language processing: Ethical considerations. In Proceedings of the First ACL Workshop on Ethics in Natural Language Processing, pages 1–11, Valencia, Spain. Association for Computational Linguistics. find
  19. Anne Lauscher and Goran Glavaš. 2019. Are we consistently biased? multidimensional analysis of biases in distributional word vectors. In *SEM 2019. find
  20. Paul Pu Liang, Yao Chong Lim, Yao-Hung Hubert Tsai, Ruslan Salakhutdinov, and Louis-Philippe Morency. 2019. Strong and simple baselines for multimodal utterance embeddings. In NAACL-HLT, pages 2599–2609, Minneapolis, Minnesota. Association for Computational Linguistics. find
  21. Paul Pu Liang, Ziyin Liu, AmirAli Bagher Zadeh, and Louis-Philippe Morency. 2018. Multimodal language analysis with recurrent multistage fusion. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. find
  22. 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
  23. Laurens van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-SNE. Journal of Machine Learning Research, 9:2579–2605. find
  24. Thomas Manzini, Lim Yao Chong, Alan W Black, and Yulia Tsvetkov. 2019. Black is to criminal as caucasian is to police: Detecting and removing multiclass bias in word embeddings. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Huma… find
  25. Chandler May, Alex Wang, Shikha Bordia, Samuel R. Bowman, and Rachel Rudinger. 2019. On measuring social biases in sentence encoders. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and… find our card
  26. Stephen Merity, Caiming Xiong, James Bradbury, and Richard Socher. 2017a. Pointer sentinel mixture models. In 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings. OpenReview.net. find our card
  27. Stephen Merity, Caiming Xiong, James Bradbury, and Richard Socher. 2017b. Pointer sentinel mixture models. In 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings. OpenReview.net. find our card
  28. Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. 2013. Efficient estimation of word representations in vector space. In 1st International Conference on Learning Representations, ICLR 2013, Scottsdale, Arizona, USA, May 2-4, 2013, Workshop Track Proceedings. find
  29. Ji Ho Park, Jamin Shin, and Pascale Fung. 2018. Reducing gender bias in abusive language detection. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 2799–2804, Brussels, Belgium. Association for Computational Linguistics. find our card
  30. Sunghyun Park, Han Suk Shim, Moitreya Chatterjee, Kenji Sagae, and Louis-Philippe Morency. 2014. Computational analysis of persuasiveness in social multimedia: A novel dataset and multimodal prediction approach. In ICMI. find
  31. Jeffrey Pennington, Richard Socher, and Christopher Manning. 2014. Glove: Global vectors for word representation. In EMNLP. find
  32. Matthew Peters, Mark Neumann, Mohit Iyyer, Matt Gardner, Christopher Clark, Kenton Lee, and Luke Zettlemoyer. 2018. Deep contextualized word representations. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technolog… find
  33. Soujanya Poria, Devamanyu Hazarika, Navonil Majumder, Gautam Naik, Erik Cambria, and Rada Mihalcea. 2019. MELD: A multimodal multi-party dataset for emotion recognition in conversations. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 527– 536, Flore… find our card
  34. Alec Radford, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. find
  35. Pranav Rajpurkar, Jian Zhang, Konstantin Lopyrev, and Percy Liang. 2016. SQuAD: 100,000+ questions for machine comprehension of text. In EMNLP. find our card
  36. Richard Socher, Alex Perelygin, Jean Wu, Jason Chuang, Christopher D. Manning, Andrew Ng, and Christopher Potts. 2013. Recursive deep models for semantic compositionality over a sentiment treebank. In EMNLP. find
  37. Chen Sun, Austin Oliver Myers, Carl Vondrick, Kevin Murphy, and Cordelia Schmid. 2019a. Videobert: A joint model for video and language representation learning. CoRR, abs/1904.01766. find our card
  38. Tony Sun, Andrew Gaut, Shirlyn Tang, Yuxin Huang, Mai ElSherief, Jieyu Zhao, Diba Mirza, Elizabeth Belding, Kai-Wei Chang, and William Yang Wang. 2019b. Mitigating gender bias in natural language processing: Literature review. In ACL. find our card
  39. Nathaniel Swinger, Maria De-Arteaga, Neil Thomas Heffernan IV, Mark DM Leiserson, and Adam Tauman Kalai. 2019. What are the biases in my word embedding? In Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society, AIES ’19, page 305–311, New York, NY, USA. Association for Computing Mac… find
  40. Jack Urbanek, Angela Fan, Siddharth Karamcheti, Saachi Jain, Samuel Humeau, Emily Dinan, Tim Rocktaschel, Douwe Kiela, Arthur Szlam, and Jason Weston. 2019. Learning to speak and act in a fantasy text adventure game. CoRR, abs/1903.03094. find
  41. Sumithra Velupillai, Hanna Suominen, Maria Liakata, Angus Roberts, Anoop D. Shah, Katherine Morley, David Osborn, Joseph Hayes, Robert Stewart, Johnny Downs, Wendy Chapman, and Rina Dutta. 2018. Using clinical natural language processing for health outcomes research: Overview and actionable suggesti… find
  42. Alex Wang, Amanpreet Singh, Julian Michael, Felix Hill, Omer Levy, and Samuel Bowman. 2018. GLUE: A multi-task benchmark and analysis platform for natural language understanding. In BlackboxNLP. find our card
  43. Chenguang Wang, Mu Li, and Alexander J. Smola. 2019. Language models with transformers. CoRR, abs/1904.09408. — [arXiv:1904.09408](https://arxiv.org/abs/1904.09408) find
  44. Alex Warstadt, Amanpreet Singh, and Samuel R Bowman. 2018. Neural network acceptability judgments. arXiv preprint arXiv:1805.12471. find
  45. Shijie Wu and Mark Dredze. 2019. Beto, bentz, becas: The surprising cross-lingual effectiveness of BERT. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 833–844,… find our card
  46. Rowan Zellers, Yonatan Bisk, Ali Farhadi, and Yejin Choi. 2019. From recognition to cognition: Visual commonsense reasoning. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR). find
  47. Brian Hu Zhang, Blake Lemoine, and Margaret Mitchell. 2018. Mitigating unwanted biases with adversarial learning. In AIES. find
  48. Jieyu Zhao, Tianlu Wang, Mark Yatskar, Ryan Cotterell, Vicente Ordonez, and Kai-Wei Chang. 2019. Gender bias in contextualized word embeddings. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1… find our card
  49. Jieyu Zhao, Tianlu Wang, Mark Yatskar, Vicente Ordonez, and Kai-Wei Chang. 2017. Men also like shopping: Reducing gender bias amplification using corpus-level constraints. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 2979–2989, Copenhagen, Denmark.… find our card
  50. Jieyu Zhao, Yichao Zhou, Zeyu Li, Wei Wang, and Kai-Wei Chang. 2018. Learning gender-neutral word embeddings. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 4847–4853, Brussels, Belgium. Association for Computational Linguistics. find our card
  51. Pei Zhou, Weijia Shi, Jieyu Zhao, Kuan-Hao Huang, Muhao Chen, Ryan Cotterell, and Kai-Wei Chang. 2019. Examining gender bias in languages with grammatical gender. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on N… find