Evaluating the Fairness of Neural Collapse in Medical Image Classification
This addresses fairness issues in medical imaging for equitable healthcare, but it is incremental as it explores an existing phenomenon (NC) in a new context.
The study investigated how Neural Collapse (NC) affects fairness in medical image classification, finding that biased training leads to different NC configurations across subgroups and a significant drop in F1 scores across all subgroups in biased settings.
Deep learning has achieved impressive performance across various medical imaging tasks. However, its inherent bias against specific groups hinders its clinical applicability in equitable healthcare systems. A recently discovered phenomenon, Neural Collapse (NC), has shown potential in improving the generalization of state-of-the-art deep learning models. Nonetheless, its implications on bias in medical imaging remain unexplored. Our study investigates deep learning fairness through the lens of NC. We analyze the training dynamics of models as they approach NC when training using biased datasets, and examine the subsequent impact on test performance, specifically focusing on label bias. We find that biased training initially results in different NC configurations across subgroups, before converging to a final NC solution by memorizing all data samples. Through extensive experiments on three medical imaging datasets -- PAPILA, HAM10000, and CheXpert -- we find that in biased settings, NC can lead to a significant drop in F1 score across all subgroups. Our code is available at https://gitlab.com/radiology/neuro/neural-collapse-fairness