Doubly-Regressing Approach for Subgroup Fairness

2025-10-23

AI Summary

Multi-sensitive-attribute settings (e.g., gender, race, age) induce combinatorial explosion of subgroups, leading to prohibitive computational overhead and severe data sparsity—critically hindering subgroup fairness evaluation and optimization. To address this, we propose a dual-regression adversarial learning framework. Our key contributions are: (1) introducing *subgroup subset fairness*, which relaxes the requirement of enumerating all exponentially many subgroups; (2) adopting the *supremum Integral Probability Metric* (supIPM) as a distributional fairness measure and deriving its differentiable upper-bound surrogate loss, enabling efficient, provably convergent optimization; and (3) jointly modeling the primary task and marginal distributions of sensitive attributes via dual regression, thereby improving generalization for underrepresented subgroups. Extensive experiments on multiple benchmark datasets demonstrate that our method significantly outperforms state-of-the-art baselines in multi-sensitive-attribute and small-subgroup regimes.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Multiagent Systems: Adversarial Agents
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

Algorithmic fairness is a socially crucial topic in real-world applications of AI. Among many notions of fairness, subgroup fairness is widely studied when multiple sensitive attributes (e.g., gender, race, age) are present. However, as the number of sensitive attributes grows, the number of subgroups increases accordingly, creating heavy computational burdens and data sparsity problem (subgroups with too small sizes). In this paper, we develop a novel learning algorithm for subgroup fairness which resolves these issues by focusing on subgroups with sufficient sample sizes as well as marginal fairness (fairness for each sensitive attribute). To this end, we formalize a notion of subgroup-subset fairness and introduce a corresponding distributional fairness measure called the supremum Integral Probability Metric (supIPM). Building on this formulation, we propose the Doubly Regressing Adversarial learning for subgroup Fairness (DRAF) algorithm, which reduces a surrogate fairness gap for supIPM with much less computation than directly reducing supIPM. Theoretically, we prove that the proposed surrogate fairness gap is an upper bound of supIPM. Empirically, we show that the DRAF algorithm outperforms baseline methods in benchmark datasets, specifically when the number of sensitive attributes is large so that many subgroups are very small.

Problem

• Addresses subgroup fairness with multiple sensitive attributes
• Solves computational burden from exponentially growing subgroups
• Mitigates data sparsity in small-sized subgroup populations

Innovation

• Doubly regressing adversarial learning for subgroup fairness
• Focusing on sufficient sample sizes and marginal fairness
• Reducing surrogate fairness gap with less computation