Enhancing Robust Fairness via Confusional Spectral Regularization

2025-01-22

AI Summary

This paper addresses the “robust fairness” problem in deep neural networks, wherein significant inter-class disparities in robust accuracy undermine equitable robustness across classes. To mitigate this, we propose an implicit optimization method based on spectral-norm regularization of the robust confusion matrix. Unlike explicit reweighting strategies, our approach is the first to derive a generalization upper bound for the worst-class robust error within the PAC-Bayesian framework, and it employs the spectral norm of the robust confusion matrix as a regularizer to alleviate distributional shift in class-wise robust performance between training and test phases. Experiments across multiple datasets and architectures demonstrate that our method substantially improves worst-class robust accuracy, enhances robust fairness, and achieves superior generalization stability compared to existing reweighting-based approaches.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Computer Vision: Adversarial Attacks & Robustness
• Natural Language Processing: Safety and Robustness

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Recent research has highlighted a critical issue known as ``robust fairness", where robust accuracy varies significantly across different classes, undermining the reliability of deep neural networks (DNNs). A common approach to address this has been to dynamically reweight classes during training, giving more weight to those with lower empirical robust performance. However, we find there is a divergence of class-wise robust performance between training set and testing set, which limits the effectiveness of these explicit reweighting methods, indicating the need for a principled alternative. In this work, we derive a robust generalization bound for the worst-class robust error within the PAC-Bayesian framework, accounting for unknown data distributions. Our analysis shows that the worst-class robust error is influenced by two main factors: the spectral norm of the empirical robust confusion matrix and the information embedded in the model and training set. While the latter has been extensively studied, we propose a novel regularization technique targeting the spectral norm of the robust confusion matrix to improve worst-class robust accuracy and enhance robust fairness. We validate our approach through comprehensive experiments on various datasets and models, demonstrating its effectiveness in enhancing robust fairness.

Problem

• robust fairness
• deep neural networks
• class-wise robust performance

Innovation

• Confusion Spectral Regularization
• robust fairness
• deep neural networks