EvoMU: Evolutionary Machine Unlearning

2026-02-02

AI Summary

This work addresses the challenge of machine unlearning—efficiently removing the influence of specific training data, such as sensitive or copyrighted content—from trained models. Existing approaches rely on manually designed, generic loss functions that often fail to adapt to diverse data distributions, leading to under- or over-unlearning. To overcome this limitation, we propose the first framework that integrates evolutionary algorithms to automatically discover task-specific unlearning loss functions without human intervention. Evaluated on the Qwen3-4B-Thinking small language model across multiple benchmarks—including TOFU-5%, TOFU-10%, MUSE, and WMDP—our method consistently outperforms current loss-based unlearning techniques. The evolved loss functions demonstrate superior efficacy and adaptability, advancing automated scientific discovery by AI co-scientists in resource-constrained settings.

Technology Category

• Machine Learning: Evolutionary Learning
• Search and Optimization: Learning to Search
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

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.

Problem

• machine unlearning
• unlearning loss
• forgetting
• retain utility
• optimal loss function

Innovation

• machine unlearning
• evolutionary search
• automatic loss discovery
• AI co-scientist
• task-specific optimization