EvoMU: Evolutionary Machine Unlearning

TL;DR

EvoMU employs evolutionary search to automatically discover task-specific unlearning loss functions, outperforming existing methods on multiple datasets with limited computational resources, advancing the field of machine unlearning.

Contribution

This work introduces an evolutionary search approach to automatically find optimal unlearning loss functions tailored to specific datasets, eliminating the need for human-designed losses.

Findings

Surpasses previous loss-based unlearning methods on multiple benchmarks

Uses a small 4B parameter model to achieve state-of-the-art results

Demonstrates the effectiveness of automatic loss discovery in machine unlearning

Abstract

Machine unlearning aims to unlearn specified training data (e.g. sensitive or copyrighted material). A prominent approach is to fine-tune an existing model with an unlearning loss that retains overall utility. The space of suitable unlearning loss functions is vast, making the search for an optimal loss function daunting. Additionally, there might not even exist a universally optimal loss function: differences in the structure and overlap of the forget and retain data can cause a loss to work well in one setting but over-unlearn or under-unlearn in another. Our approach EvoMU tackles these two challenges simultaneously. An evolutionary search procedure automatically finds task-specific losses in the vast space of possible unlearning loss functions. This allows us to find dataset-specific losses that match or outperform existing losses from the literature, without the need for a human-in-the-loop. This work is therefore an instance of automatic scientific discovery, a.k.a. an AI co-scientist. In contrast to previous AI co-scientist works, we do so on a budget: We achieve SotA results using a small 4B parameter model (Qwen3-4B-Thinking), showing the potential of AI co-scientists with limited computational resources. Our experimental evaluation shows that we surpass previous loss-based unlearning formulations on TOFU-5%, TOFU-10%, MUSE and WMDP by synthesizing novel unlearning losses. Our code is available at https://github.com/Batorskq/EvoMU.