Are Gender-Neutral Queries Really Gender-Neutral? Mitigating Gender Bias in Image Search

TL;DR

This paper investigates gender bias in image search results for gender-neutral queries, diagnosing existing models and proposing novel debiasing methods that significantly reduce bias in benchmark datasets.

Contribution

The paper introduces two innovative debiasing techniques—an in-processing fair sampling method and a post-processing feature clipping approach—applying them to both specialized and pre-trained models.

Findings

Both models exhibit severe gender bias.

Proposed methods effectively reduce gender bias.

Significant bias reduction demonstrated on MS-COCO and Flickr30K.

Abstract

Internet search affects people's cognition of the world, so mitigating biases in search results and learning fair models is imperative for social good. We study a unique gender bias in image search in this work: the search images are often gender-imbalanced for gender-neutral natural language queries. We diagnose two typical image search models, the specialized model trained on in-domain datasets and the generalized representation model pre-trained on massive image and text data across the internet. Both models suffer from severe gender bias. Therefore, we introduce two novel debiasing approaches: an in-processing fair sampling method to address the gender imbalance issue for training models, and a post-processing feature clipping method base on mutual information to debias multimodal representations of pre-trained models. Extensive experiments on MS-COCO and Flickr30K benchmarks show that our methods significantly reduce the gender bias in image search models.