{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2023,11,22]],"date-time":"2023-11-22T06:49:11Z","timestamp":1700635751255},"reference-count":49,"publisher":"MIT Press","license":[{"start":{"date-parts":[[2022,10,20]],"date-time":"2022-10-20T00:00:00Z","timestamp":1666224000000},"content-version":"vor","delay-in-days":292,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["direct.mit.edu"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2022,10,18]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Text representations learned by machine learning models often encode undesirable demographic information of the user. Predictive models based on these representations can rely on such information, resulting in biased decisions. We present a novel debiasing technique, Fairness-aware Rate Maximization (FaRM), that removes protected information by making representations of instances belonging to the same protected attribute class uncorrelated, using the rate-distortion function. FaRM is able to debias representations with or without a target task at hand. FaRM can also be adapted to remove information about multiple protected attributes simultaneously. Empirical evaluations show that FaRM achieves state-of-the-art performance on several datasets, and learned representations leak significantly less protected attribute information against an attack by a non-linear probing network.<\/jats:p>","DOI":"10.1162\/tacl_a_00512","type":"journal-article","created":{"date-parts":[[2022,10,20]],"date-time":"2022-10-20T14:52:41Z","timestamp":1666277561000},"page":"1159-1174","update-policy":"http:\/\/dx.doi.org\/10.1162\/mitpressjournals.corrections.policy","source":"Crossref","is-referenced-by-count":1,"title":["Learning Fair Representations via Rate-Distortion Maximization"],"prefix":"10.1162","volume":"10","author":[{"given":"Somnath Basu Roy","family":"Chowdhury","sequence":"first","affiliation":[{"name":"UNC Chapel Hill, USA somnath@cs.unc.edu"}]},{"given":"Snigdha","family":"Chaturvedi","sequence":"additional","affiliation":[{"name":"UNC Chapel Hill, USA snigdha@cs.unc.edu"}]}],"member":"281","published-online":{"date-parts":[[2022,10,18]]},"reference":[{"key":"2022102014520019100_bib1","doi-asserted-by":"publisher","first-page":"19","DOI":"10.3115\/1620754.1620758","article-title":"A study on similarity and relatedness using distributional and WordNet-based approaches","volume-title":"Proceedings of Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics","author":"Agirre","year":"2009"},{"key":"2022102014520019100_bib2","article-title":"Layer normalization","author":"Ba","year":"2016","journal-title":"arXiv preprint arXiv:1607.06450"},{"key":"2022102014520019100_bib3","doi-asserted-by":"publisher","first-page":"6330","DOI":"10.18653\/v1\/D19-1662","article-title":"Adversarial removal of demographic attributes revisited","volume-title":"Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)","author":"Barrett","year":"2019"},{"issue":"2","key":"2022102014520019100_bib4","doi-asserted-by":"publisher","first-page":"65","DOI":"10.3390\/data4020065","article-title":"Predictive models of student college commitment decisions using machine learning","volume":"4","author":"Basu","year":"2019","journal-title":"Data"},{"key":"2022102014520019100_bib5","doi-asserted-by":"publisher","first-page":"550","DOI":"10.18653\/v1\/2021.emnlp-main.43","article-title":"Adversarial scrubbing of demographic information for text classification","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"Roy Chowdhury","year":"2021"},{"issue":"4175","key":"2022102014520019100_bib6","doi-asserted-by":"publisher","first-page":"398","DOI":"10.1126\/science.187.4175.398","article-title":"Sex bias in graduate admissions: Data from berkeley: Measuring bias is harder than is usually assumed, and the evidence is sometimes contrary to expectation","volume":"187","author":"Bickel","year":"1975","journal-title":"Science"},{"key":"2022102014520019100_bib7","doi-asserted-by":"publisher","first-page":"1119","DOI":"10.18653\/v1\/D16-1120","article-title":"Demographic dialectal variation in social media: A case study of African-American English","volume-title":"Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing","author":"Lin Blodgett","year":"2016"},{"key":"2022102014520019100_bib8","first-page":"4349","article-title":"Man is to computer programmer as woman is to homemaker? debiasing word embeddings","volume-title":"Advances in Neural Information Processing Systems 29: Annual Conference on Neural Information Processing Systems 2016, December 5\u201310, 2016, Barcelona, Spain","author":"Bolukbasi","year":"2016"},{"key":"2022102014520019100_bib9","first-page":"1301","article-title":"Discriminating gender on Twitter","volume-title":"Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing","author":"Burger","year":"2011"},{"issue":"114","key":"2022102014520019100_bib10","first-page":"1","article-title":"Redunet: A white-box deep network from the principle of maximizing rate reduction","volume":"23","author":"Chan","year":"2022","journal-title":"Journal of Machine Learning Research"},{"key":"2022102014520019100_bib11","volume-title":"Elements of Information Theory","author":"Cover","year":"1999"},{"key":"2022102014520019100_bib12","doi-asserted-by":"publisher","first-page":"120","DOI":"10.1145\/3287560.3287572","article-title":"Bias in bios: A case study of semantic representation bias in a high-stakes setting","volume-title":"proceedings of the Conference on Fairness, Accountability, and Transparency","author":"De-Arteaga","year":"2019"},{"key":"2022102014520019100_bib13","doi-asserted-by":"publisher","first-page":"5034","DOI":"10.18653\/v1\/2021.emnlp-main.411","article-title":"OSCaR: Orthogonal subspace correction and rectification of biases in word embeddings","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"Dev","year":"2021"},{"key":"2022102014520019100_bib14","first-page":"4171","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","volume-title":"Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers)","author":"Devlin","year":"2019"},{"key":"2022102014520019100_bib15","doi-asserted-by":"publisher","first-page":"11","DOI":"10.18653\/v1\/D18-1002","article-title":"Adversarial removal of demographic attributes from text data","volume-title":"Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing","author":"Elazar","year":"2018"},{"key":"2022102014520019100_bib16","doi-asserted-by":"publisher","first-page":"160","DOI":"10.1162\/tacl_a_00359","article-title":"Amnesic probing: Behavioral explanation with amnesic counterfactuals","volume":"9","author":"Elazar","year":"2021","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2022102014520019100_bib17","doi-asserted-by":"publisher","first-page":"1615","DOI":"10.18653\/v1\/D17-1169","article-title":"Using millions of emoji occurrences to learn any-domain representations for detecting sentiment, emotion and sarcasm","volume-title":"Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing","author":"Felbo","year":"2017"},{"key":"2022102014520019100_bib18","volume-title":"Function Spaces","author":"Fleming","year":"2003"},{"issue":"4","key":"2022102014520019100_bib19","doi-asserted-by":"publisher","DOI":"10.14569\/IJACSA.2016.070467","article-title":"Improving credit scorecard modeling through applying text analysis","volume":"7","author":"Ghailan","year":"2016","journal-title":"Institutions"},{"key":"2022102014520019100_bib20","article-title":"Generative adversarial nets","volume-title":"Advances in Neural Information Processing Systems","author":"Goodfellow","year":"2014"},{"key":"2022102014520019100_bib21","doi-asserted-by":"publisher","first-page":"1406","DOI":"10.1145\/2339530.2339751","article-title":"Large-scale learning of word relatedness with constraints","volume-title":"The 18th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD \u201912, Beijing, China, August 12\u201316, 2012","author":"Halawi","year":"2012"},{"issue":"4","key":"2022102014520019100_bib22","doi-asserted-by":"publisher","first-page":"665","DOI":"10.1162\/COLI_a_00237","article-title":"SimLex-999: Evaluating semantic models with (genuine) similarity estimation","volume":"41","author":"Hill","year":"2015","journal-title":"Computational Linguistics"},{"key":"2022102014520019100_bib23","doi-asserted-by":"publisher","first-page":"427","DOI":"10.18653\/v1\/E17-2068","article-title":"Bag of tricks for efficient text classification","volume-title":"Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers","author":"Joulin","year":"2017"},{"issue":"4","key":"2022102014520019100_bib24","doi-asserted-by":"publisher","first-page":"401","DOI":"10.1093\/llc\/17.4.401","article-title":"Automatically categorizing written texts by author gender","volume":"17","author":"Koppel","year":"2002","journal-title":"Literary and Linguistic Computing"},{"key":"2022102014520019100_bib25","doi-asserted-by":"crossref","first-page":"25","DOI":"10.18653\/v1\/P18-2005","article-title":"Towards robust and privacy-preserving text representations","volume-title":"Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)","author":"Li","year":"2018"},{"key":"2022102014520019100_bib26","article-title":"Decoupled weight decay regularization","volume-title":"7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6\u20139, 2019","author":"Loshchilov","year":"2019"},{"issue":"9","key":"2022102014520019100_bib27","doi-asserted-by":"publisher","first-page":"1546","DOI":"10.1109\/TPAMI.2007.1085","article-title":"Segmentation of multivariate mixed data via lossy data coding and compression","volume":"29","author":"Yi","year":"2007","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"2022102014520019100_bib28","first-page":"142","article-title":"Learning word vectors for sentiment analysis","volume-title":"Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies","author":"Maas","year":"2011"},{"issue":"11","key":"2022102014520019100_bib29","article-title":"Visualizing data using t-sne","volume":"9","author":"Van der Maaten","year":"2008","journal-title":"Journal of Machine Learning Research"},{"key":"2022102014520019100_bib30","article-title":"A rate- distortion framework for explaining neural network decisions","author":"Macdonald","year":"2019","journal-title":"ArXiv preprint"},{"issue":"6","key":"2022102014520019100_bib31","doi-asserted-by":"publisher","DOI":"10.1145\/3457607","article-title":"A survey on bias and fairness in machine learning","volume":"54","author":"Mehrabi","year":"2021","journal-title":"ACM Computing Surveys,"},{"key":"2022102014520019100_bib32","article-title":"\u201cHow old do you think i am?\u201d A study of language and age in Twitter","volume-title":"Proceedings of the International AAAI Conference on Web and Social Media","author":"Nguyen","year":"2013"},{"key":"2022102014520019100_bib33","first-page":"2825","article-title":"Scikit- learn: Machine learning in Python","volume":"12","author":"Pedregosa","year":"2011","journal-title":"Journal of Machine Learning Research"},{"key":"2022102014520019100_bib34","first-page":"2089","article-title":"A universal part-of-speech tagset","volume-title":"Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC\u201912)","author":"Petrov","year":"2012"},{"key":"2022102014520019100_bib35","first-page":"750","article-title":"Overview of the 4th author profiling task at pan 2016: Cross-genre evaluations","volume":"2016","author":"Rangel","year":"2016","journal-title":"Working Notes Papers of the CLEF"},{"key":"2022102014520019100_bib36","doi-asserted-by":"publisher","first-page":"7237","DOI":"10.18653\/v1\/2020.acl-main.647","article-title":"Null it out: Guarding protected attributes by iterative nullspace projection","volume-title":"Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics","author":"Ravfogel","year":"2020"},{"key":"2022102014520019100_bib37","doi-asserted-by":"publisher","first-page":"4187","DOI":"10.18653\/v1\/N19-1424","article-title":"What\u2019s in a name? Reducing bias in bios without access to protected attributes","volume-title":"Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers)","author":"Romanov","year":"2019"},{"key":"2022102014520019100_bib38","first-page":"410","article-title":"V-measure: A conditional entropy-based external cluster evaluation measure","volume-title":"Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL)","author":"Rosenberg","year":"2007"},{"key":"2022102014520019100_bib39","doi-asserted-by":"publisher","first-page":"2492","DOI":"10.18653\/v1\/2021.emnlp-main.193","article-title":"Evaluating debiasing techniques for intersectional biases","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"Subramanian","year":"2021"},{"key":"2022102014520019100_bib40","first-page":"3081","article-title":"CLiPS stylometry investigation (CSI) corpus: A Dutch corpus for the detection of age, gender, personality, sentiment and deception in text","volume-title":"Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC\u201914)","author":"Verhoeven","year":"2014"},{"key":"2022102014520019100_bib41","first-page":"1632","article-title":"TwiSty: A multilingual Twitter stylometry corpus for gender and personality profiling","volume-title":"Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC\u201916)","author":"Verhoeven","year":"2016"},{"key":"2022102014520019100_bib42","doi-asserted-by":"publisher","first-page":"183","DOI":"10.18653\/v1\/2020.emnlp-main.14","article-title":"Information-theoretic probing with minimum description length","volume-title":"Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)","author":"Voita","year":"2020"},{"issue":"3","key":"2022102014520019100_bib43","first-page":"266","article-title":"Examining multiple features for author profiling","volume":"5","author":"Weren","year":"2014","journal-title":"Journal of Information and Data Management"},{"key":"2022102014520019100_bib44","first-page":"585","article-title":"Controllable invariance through adversarial feature learning","volume-title":"Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, December 4\u20139, 2017, Long Beach, CA, USA","author":"Xie","year":"2017"},{"key":"2022102014520019100_bib45","first-page":"9422","article-title":"Learning diverse and discriminative representations via the principle of maximal coding rate reduction","volume-title":"Advances in Neural Information Processing Systems","author":"Yaodong","year":"2020"},{"key":"2022102014520019100_bib46","first-page":"325","article-title":"Learning fair representations","volume-title":"Proceedings of the 30th International Conference on Machine Learning, ICML 2013, Atlanta, GA, USA, 16\u201321 June 2013","author":"Zemel","year":"2013"},{"key":"2022102014520019100_bib47","doi-asserted-by":"publisher","first-page":"335","DOI":"10.1145\/3278721.3278779","article-title":"Mitigating unwanted biases with adversarial learning","volume-title":"Proceedings of the 2018 AAAI\/ACM Conference on AI, Ethics, and Society","author":"Zhang","year":"2018"},{"key":"2022102014520019100_bib48","first-page":"15649","article-title":"Inherent tradeoffs in learning fair representations","volume-title":"Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019, NeurIPS 2019, December 8\u201314, 2019, Vancouver, BC, Canada","author":"Zhao","year":"2019"},{"key":"2022102014520019100_bib49","doi-asserted-by":"publisher","first-page":"4847","DOI":"10.18653\/v1\/D18-1521","article-title":"Learning gender-neutral word embeddings","volume-title":"Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing","author":"Zhao","year":"2018"}],"container-title":["Transactions of the Association for Computational Linguistics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/direct.mit.edu\/tacl\/article-pdf\/doi\/10.1162\/tacl_a_00512\/2054697\/tacl_a_00512.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/direct.mit.edu\/tacl\/article-pdf\/doi\/10.1162\/tacl_a_00512\/2054697\/tacl_a_00512.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,10,20]],"date-time":"2022-10-20T14:53:45Z","timestamp":1666277625000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/tacl\/article\/doi\/10.1162\/tacl_a_00512\/113492\/Learning-Fair-Representations-via-Rate-Distortion"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022]]},"references-count":49,"URL":"https:\/\/doi.org\/10.1162\/tacl_a_00512","relation":{},"ISSN":["2307-387X"],"issn-type":[{"value":"2307-387X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2022]]},"published":{"date-parts":[[2022]]}}}