From Gradient Volume to Shapley Fairness: Towards Fair Multi-Task Learning

Xiao Wang, Yuying Han, Dazi Li, Fei Zhang, Min Tang

Keywords: Multi-task Learning, Shapley Value, Fair Optimization

TL;DR: We introduce SVFair, a fairness-driven method using Shapley values to quantify and mitigate gradient conflicts in multi-task learning, achieving state-of-the-art performance through the novel Volume Determinant (VolDet and VolDetPro) metric.

Abstract: Multi-task learning often suffers from gradient conflicts, leading to unfair optimization and degraded overall performance. To address this, we present SVFair, a Shapley value-based framework for fair gradient aggregation. We propose two scalable geometric conflict metrics: VolDet, a gram determinant volume metric, and VolDetPro, its sign-aware extension distinguishing antagonistic gradients. By integrating these metrics into Shapley value computation, SVFair quantifies each task’s deviation from the overall gradient and rebalances updates toward fairness. In parallel, our Shapley value computation admits controllable complexity. Extensive experiments show that SVFair achieves state-of-the-art results across diverse supervised and reinforcement learning benchmarks, and further improves existing methods when integrated as a fairness-enhancing module.

Primary Area: alignment, fairness, safety, privacy, and societal considerations

Submission Number: 5136