MUSO: Achieving Exact Machine Unlearning in Over-Parameterized Regimes

TL;DR

This paper introduces MUSO, a method for achieving exact machine unlearning in over-parameterized models, using analytical frameworks and algorithms that outperform existing relabeling-based approaches.

Contribution

It provides the first theoretical demonstration of exact MU in over-parameterized linear models and proposes a practical algorithm for nonlinear networks.

Findings

Exact MU achieved in over-parameterized linear models.

Proposed algorithm unifies unlearning and relabeling tasks.

Numerical experiments show superior performance over state-of-the-art methods.

Abstract

Machine unlearning (MU) is to make a well-trained model behave as if it had never been trained on specific data. In today's over-parameterized models, dominated by neural networks, a common approach is to manually relabel data and fine-tune the well-trained model. It can approximate the MU model in the output space, but the question remains whether it can achieve exact MU, i.e., in the parameter space. We answer this question by employing random feature techniques to construct an analytical framework. Under the premise of model optimization via stochastic gradient descent, we theoretically demonstrated that over-parameterized linear models can achieve exact MU through relabeling specific data. We also extend this work to real-world nonlinear networks and propose an alternating optimization algorithm that unifies the tasks of unlearning and relabeling. The algorithm's effectiveness, confirmed through numerical experiments, highlights its superior performance in unlearning across various scenarios compared to current state-of-the-art methods, particularly excelling over similar relabeling-based MU approaches.