FairQuant: Fairness-Aware Mixed-Precision Quantization for Medical Image Classification

2026-02-26

AI Summary

This work addresses the issue of algorithmic fairness in medical image classification under quantization, where conventional approaches often overlook performance disparities across demographic groups. To this end, we propose FairQuant, a novel framework that explicitly integrates fairness into mixed-precision quantization. FairQuant employs group-aware importance analysis, budget-constrained bit allocation, and a learnable bit-aware quantization (BAQ) mechanism, further enhanced by a fairness regularization term that jointly optimizes model weights and per-layer bit assignments. Evaluated on the Fitzpatrick17k and ISIC2019 datasets, FairQuant configurations using 4–6 bits achieve accuracy comparable to uniform 8-bit quantization while significantly improving performance for the worst-performing subgroups. Moreover, under the same bit budget, FairQuant consistently outperforms uniform 4-bit and 8-bit baselines in fairness metrics.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Machine Learning: Calibration & Uncertainty Quantification
• Philosophy and Ethics of AI: Bias, Fairness & Equity

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments
• Social Networks and Social Media: Fairness and bias in social network and social media analysis

Abstract

Compressing neural networks by quantizing model parameters offers useful trade-off between performance and efficiency. Methods like quantization-aware training and post-training quantization strive to maintain the downstream performance of compressed models compared to the full precision models. However, these techniques do not explicitly consider the impact on algorithmic fairness. In this work, we study fairness-aware mixed-precision quantization schemes for medical image classification under explicit bit budgets. We introduce FairQuant, a framework that combines group-aware importance analysis, budgeted mixed-precision allocation, and a learnable Bit-Aware Quantization (BAQ) mode that jointly optimizes weights and per-unit bit allocations under bitrate and fairness regularization. We evaluate the method on Fitzpatrick17k and ISIC2019 across ResNet18/50, DeiT-Tiny, and TinyViT. Results show that FairQuant configurations with average precision near 4-6 bits recover much of the Uniform 8-bit accuracy while improving worst-group performance relative to Uniform 4- and 8-bit baselines, with comparable fairness metrics under shared budgets.

Problem

• fairness-aware quantization
• mixed-precision quantization
• medical image classification
• algorithmic fairness
• model compression

Innovation

• fairness-aware quantization
• mixed-precision quantization
• bit-aware quantization
• algorithmic fairness
• medical image classification