A Unified Gradient-based Framework for Task-agnostic Continual Learning-Unlearning

2025-05-21

AI Summary

This work addresses the fundamental challenge of jointly optimizing continual learning (CL) and machine unlearning (MU), proposing a unified Continual Learning–Unlearning (CLU) framework that simultaneously ensures plasticity for knowledge updating and stability for retaining historical knowledge. Methodologically, it first uncovers the intrinsic gradient-optimization unity between CL and MU, introducing a remain-preserved manifold constraint and a fast-slow weight adaptation mechanism. It further designs a suite of techniques—including KL-divergence-based gradient decomposition, Hessian compensation, saliency-aware weight masking, and second-order directional approximation—enabling task-agnostic, cross-class, and sample-level fine-grained unlearning. Extensive experiments across multiple datasets and architectures demonstrate that CLU significantly improves incremental learning accuracy, unlearning precision, and knowledge stability. The framework establishes a novel paradigm and scalable methodology for building dynamically compliant AI systems.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Search and Optimization: Learning to Search
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• 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

Abstract

Recent advancements in deep models have highlighted the need for intelligent systems that combine continual learning (CL) for knowledge acquisition with machine unlearning (MU) for data removal, forming the Continual Learning-Unlearning (CLU) paradigm. While existing work treats CL and MU as separate processes, we reveal their intrinsic connection through a unified optimization framework based on Kullback-Leibler divergence minimization. This framework decomposes gradient updates for approximate CLU into four components: learning new knowledge, unlearning targeted data, preserving existing knowledge, and modulation via weight saliency. A critical challenge lies in balancing knowledge update and retention during sequential learning-unlearning cycles. To resolve this stability-plasticity dilemma, we introduce a remain-preserved manifold constraint to induce a remaining Hessian compensation for CLU iterations. A fast-slow weight adaptation mechanism is designed to efficiently approximate the second-order optimization direction, combined with adaptive weighting coefficients and a balanced weight saliency mask, proposing a unified implementation framework for gradient-based CLU. Furthermore, we pioneer task-agnostic CLU scenarios that support fine-grained unlearning at the cross-task category and random sample levels beyond the traditional task-aware setups. Experiments demonstrate that the proposed UG-CLU framework effectively coordinates incremental learning, precise unlearning, and knowledge stability across multiple datasets and model architectures, providing a theoretical foundation and methodological support for dynamic, compliant intelligent systems.

Problem

• Unifying continual learning and unlearning via gradient optimization
• Balancing knowledge update and retention in sequential cycles
• Enabling task-agnostic unlearning at fine-grained data levels

Innovation

• Unified gradient framework for CLU optimization
• Remain-preserved manifold with Hessian compensation
• Fast-slow weight adaptation for second-order direction