Towards Reducing Bias in Gender Classification
arXiv:1911.08556
Abstract
Societal bias towards certain communities is a big problem that affects a lot of machine learning systems. This work aims at addressing the racial bias present in many modern gender recognition systems. We learn race invariant representations of human faces with an adversarially trained autoencoder model. We show that such representations help us achieve less biased performance in gender classification. We use variance in classification accuracy across different races as a surrogate for the racial bias of the model and achieve a drop of over 40% in variance with race invariant representations.
arXiv commentarXiv admin note: text overlap with arXiv:1706.00409 by other authors
References in corpus (7)
- Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings
- Word Translation Without Parallel Data
- Unsupervised Machine Translation Using Monolingual Corpora Only
- Men Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level Constraints
- Censoring Representations with an Adversary
- Learning to Pivot with Adversarial Networks
- Examining Gender and Race Bias in Two Hundred Sentiment Analysis Systems