Optimising Equal Opportunity Fairness in Model TrainingDownload PDF

Anonymous

08 Mar 2022 (modified: 05 May 2023)NAACL 2022 Conference Blind SubmissionReaders: Everyone
Paper Link: https://openreview.net/forum?id=mtVeNh_4dh8
Paper Type: Long paper (up to eight pages of content + unlimited references and appendices)
Abstract: Real-world datasets often encode stereotypes and societal biases. Such biases can be implicitly captured by trained models, leading to biased predictions and exacerbating existing societal preconceptions. Existing debiasing methods, such as adversarial training and removing protected information from representations, have been shown to reduce bias. However, a disconnect between fairness criteria and training objectives makes it difficult to reason theoretically about the effectiveness of different techniques. In this work, we propose two novel training objectives which directly optimise for the widely-used criterion of {\it equal opportunity}, and show that they are effective in reducing bias while maintaining high performance over two classification tasks.
Presentation Mode: This paper will be presented in person in Seattle
Copyright Consent Signature (type Name Or NA If Not Transferrable): Aili Shen
Copyright Consent Name And Address: The University of Melbourne, Parkville VIC 3010, Australia
0 Replies

Loading