Erase to Retain: Low Rank Adaptation Guided Selective Unlearning in Medical Segmentation Networks

TL;DR

This paper presents Erase to Retain, a novel low-rank adaptation guided framework for controllable unlearning in medical segmentation networks, enabling targeted forgetting while maintaining overall performance.

Contribution

It introduces a LoRA-based subspace unlearning method that selectively erases specific knowledge in medical image segmentation models without full retraining.

Findings

Significantly reduces forget-set IoU in segmentation tasks

Maintains or improves performance on retained data and validation sets

Effectively decreases classification accuracy on sensitive subsets

Abstract

The ability to selectively remove knowledge from medical segmentation networks is increasingly important for privacy compliance, ethical deployment, and continual dataset revision. We introduce Erase to Retain, a controllable unlearning framework for medical image segmentation that achieves targeted forgetting without full retraining. Our method uses a teacher-student distillation paradigm with Low-Rank Adaptation (LoRA) constrained subspace updates, enabling the student network to erase lesion-specific or class-specific representations in low-rank decoder spaces while preserving global anatomical understanding. During the strong unlearning phase, LoRA modules are adversarially optimized to contradict the teacher's confident predictions on a designated forget subset, enforcing semantic removal. This is followed by a gentle restoration phase that recovers generalization on retained data through head-only supervised refinement. For ISIC segmentation, the student reduces forget-set IoU from 0.875 to 0.509 while maintaining competitive performance on the retain and validation splits (0.647 to 0.677 IoU). On the cross-domain CHASE dataset, Erase to Retain consistently lowers forget-set IoU while preserving utility on retain and validation sets. For ISIC classification, our method decreases accuracy on the forget subset from 87.0 percent to 64.1 percent while improving retain accuracy from 83.9 percent to 90.6 percent. These results demonstrate that LoRA-based subspace unlearning provides a practical pathway toward responsible, controllable, and reversible unlearning in medical image analysis, enabling models to forget sensitive samples or structures while preserving performance where it matters most.