Machine Unlearning for Medical Imaging

TL;DR

This paper evaluates the effectiveness and efficiency of machine unlearning algorithms in medical imaging, highlighting their current strengths and limitations in removing specific training data influence while maintaining model fairness.

Contribution

It provides a comprehensive evaluation of existing unlearning algorithms in medical imaging, revealing their performance, biases, and computational challenges.

Findings

Unlearning algorithms perform well on retain and forget sets.

They show no bias against gender in the evaluated datasets.

Generalization may decrease with larger forget sets.

Abstract

Machine unlearning is the process of removing the impact of a particular set of training samples from a pretrained model. It aims to fulfill the "right to be forgotten", which grants the individuals such as patients the right to reconsider their contribution in models including medical imaging models. In this study, we evaluate the effectiveness (performance) and computational efficiency of different unlearning algorithms in medical imaging domain. Our evaluations demonstrate that the considered unlearning algorithms perform well on the retain set (samples whose influence on the model is allowed to be retained) and forget set (samples whose contribution to the model should be eliminated), and show no bias against male or female samples. They, however, adversely impact the generalization of the model, especially for larger forget set sizes. Moreover, they might be biased against easy or hard samples, and need additional computational overhead for hyper-parameter tuning. In conclusion, machine unlearning seems promising for medical imaging, but the existing unlearning algorithms still needs further improvements to become more practical for medical applications.