Stride-Net: Fairness-Aware Disentangled Representation Learning for Chest X-Ray Diagnosis

2026-02-11

AI Summary

This work addresses the disparity in performance of chest X-ray diagnostic models across subgroups defined by race and gender, which poses fairness and clinical safety concerns. The authors propose a decoupled learning framework that internalizes fairness as an objective of representation learning. By integrating a learnable mask, adversarial confusion loss, and a Group Optimal Transport–based semantic alignment mechanism, the method constructs disease-discriminative representations robust to sensitive attributes. Implemented on both ResNet and Vision Transformer architectures and augmented with BioBERT label embeddings, the approach significantly improves fairness metrics across racial and intersectional gender–race subgroups on the MIMIC-CXR and CheXpert datasets, while maintaining or surpassing baseline diagnostic accuracy and outperforming existing debiasing methods.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Machine Learning: Ethics, Bias, and Fairness
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

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
• Security and Privacy: Data transparency and provenance

Abstract

Deep neural networks for chest X-ray classification achieve strong average performance, yet often underperform for specific demographic subgroups, raising critical concerns about clinical safety and equity. Existing debiasing methods frequently yield inconsistent improvements across datasets or attain fairness by degrading overall diagnostic utility, treating fairness as a post hoc constraint rather than a property of the learned representation. In this work, we propose Stride-Net (Sensitive Attribute Resilient Learning via Disentanglement and Learnable Masking with Embedding Alignment), a fairness-aware framework that learns disease-discriminative yet demographically invariant representations for chest X-ray analysis. Stride-Net operates at the patch level, using a learnable stride-based mask to select label-aligned image regions while suppressing sensitive attribute information through adversarial confusion loss. To anchor representations in clinical semantics and discourage shortcut learning, we further enforce semantic alignment between image features and BioBERT-based disease label embeddings via Group Optimal Transport. We evaluate Stride-Net on the MIMIC-CXR and CheXpert benchmarks across race and intersectional race-gender subgroups. Across architectures including ResNet and Vision Transformers, Stride-Net consistently improves fairness metrics while matching or exceeding baseline accuracy, achieving a more favorable accuracy-fairness trade-off than prior debiasing approaches. Our code is available at https://github.com/Daraksh/Fairness_StrideNet.

Problem

• fairness
• disentangled representation
• chest X-ray diagnosis
• demographic bias
• clinical equity

Innovation

• disentangled representation learning
• fairness-aware learning
• learnable masking
• semantic alignment
• adversarial debiasing