Don't Break the Boundary: Continual Unlearning for OOD Detection Based on Free Energy Repulsion

2026-02-06

AI Summary

This work addresses the challenge that conventional machine unlearning methods in open-world settings often disrupt the intrinsic manifold structure of in-distribution (ID) data, leading to catastrophic degradation in out-of-distribution (OOD) detection performance. To bridge this gap, the paper establishes a unified perspective between OOD detection and machine unlearning, proposing a novel boundary-preserving class unlearning paradigm: it reformulates the removal of a target class as its transformation into an OOD sample. The authors introduce a Push-Pull game mechanism based on free energy repulsion—low-energy attraction for retained classes and high-energy repulsion for forgotten classes—to stabilize the ID manifold during unlearning. Integrated with parameter-efficient fine-tuning, the resulting TFER framework enables efficient continual unlearning. Experiments demonstrate that the method successfully fulfills unlearning objectives while significantly preserving the model’s discriminative capability for both retained and OOD data, exhibiting superior stability in continual learning scenarios.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Deploying trustworthy AI in open-world environments faces a dual challenge: the necessity for robust Out-of-Distribution (OOD) detection to ensure system safety, and the demand for flexible machine unlearning to satisfy privacy compliance and model rectification. However, this objective encounters a fundamental geometric contradiction: current OOD detectors rely on a static and compact data manifold, whereas traditional classification-oriented unlearning methods disrupt this delicate structure, leading to a catastrophic loss of the model's capability to discriminate anomalies while erasing target classes. To resolve this dilemma, we first define the problem of boundary-preserving class unlearning and propose a pivotal conceptual shift: in the context of OOD detection, effective unlearning is mathematically equivalent to transforming the target class into OOD samples. Based on this, we propose the TFER (Total Free Energy Repulsion) framework. Inspired by the free energy principle, TFER constructs a novel Push-Pull game mechanism: it anchors retained classes within a low-energy ID manifold through a pull mechanism, while actively expelling forgotten classes to high-energy OOD regions using a free energy repulsion force. This approach is implemented via parameter-efficient fine-tuning, circumventing the prohibitive cost of full retraining. Extensive experiments demonstrate that TFER achieves precise unlearning while maximally preserving the model's discriminative performance on remaining classes and external OOD data. More importantly, our study reveals that the unique Push-Pull equilibrium of TFER endows the model with inherent structural stability, allowing it to effectively resist catastrophic forgetting without complex additional constraints, thereby demonstrating exceptional potential in continual unlearning tasks.

Problem

• continual unlearning
• Out-of-Distribution detection
• machine unlearning
• data manifold
• boundary preservation

Innovation

• Continual Unlearning
• Out-of-Distribution Detection
• Free Energy Principle
• Push-Pull Mechanism
• Boundary Preservation