Invisible Watermarks, Visible Gains: Steering Machine Unlearning with Bi-Level Watermarking Design
This addresses the need for regulatory compliance and trust in machine learning by enabling more effective removal of sensitive data influences, though it is an incremental improvement over existing unlearning methods.
The paper tackles the problem of machine unlearning by proposing a data-level approach using digital watermarking, specifically introducing Water4MU, a bi-level optimization framework that modifies data content to facilitate precise removal of sensitive data while maintaining model utility, and it outperforms existing methods in challenging unlearning scenarios.
With the increasing demand for the right to be forgotten, machine unlearning (MU) has emerged as a vital tool for enhancing trust and regulatory compliance by enabling the removal of sensitive data influences from machine learning (ML) models. However, most MU algorithms primarily rely on in-training methods to adjust model weights, with limited exploration of the benefits that data-level adjustments could bring to the unlearning process. To address this gap, we propose a novel approach that leverages digital watermarking to facilitate MU by strategically modifying data content. By integrating watermarking, we establish a controlled unlearning mechanism that enables precise removal of specified data while maintaining model utility for unrelated tasks. We first examine the impact of watermarked data on MU, finding that MU effectively generalizes to watermarked data. Building on this, we introduce an unlearning-friendly watermarking framework, termed Water4MU, to enhance unlearning effectiveness. The core of Water4MU is a bi-level optimization (BLO) framework: at the upper level, the watermarking network is optimized to minimize unlearning difficulty, while at the lower level, the model itself is trained independently of watermarking. Experimental results demonstrate that Water4MU is effective in MU across both image classification and image generation tasks. Notably, it outperforms existing methods in challenging MU scenarios, known as "challenging forgets".