FairASR: Fair Audio Contrastive Learning for Automatic Speech Recognition

TL;DR

FairASR introduces a novel contrastive learning approach with a gradient reversal layer to reduce demographic bias in speech recognition models, achieving fairer performance across groups without sacrificing accuracy.

Contribution

It presents a new fairness-oriented training method for ASR that suppresses demographic information in learned representations using adversarial contrastive learning.

Findings

Significantly reduces demographic performance disparities

Maintains competitive overall ASR accuracy

Effective bias mitigation across multiple demographic groups

Abstract

Large-scale ASR models have achieved remarkable gains in accuracy and robustness. However, fairness issues remain largely unaddressed despite their critical importance in real-world applications. In this work, we introduce FairASR, a system that mitigates demographic bias by learning representations that are uninformative about group membership, enabling fair generalization across demographic groups. Leveraging a multi-demographic dataset, our approach employs a gradient reversal layer to suppress demographic-discriminative features while maintaining the ability to capture generalizable speech patterns through an unsupervised contrastive loss. Experimental results show that FairASR delivers competitive overall ASR performance while significantly reducing performance disparities across different demographic groups.