arXiv 2503.233902025-03-30cs.LGcs.AI

Pareto Continual Learning：面向动态稳定性-可塑性权衡的偏好条件化学习与适应

Pareto Continual Learning: Preference-Conditioned Learning and Adaption for Dynamic Stability-Plasticity Trade-off

City University of Hong Kong（香港城市大学）；HK Institute of Science & Innovation, Chinese Academy of Sciences（中国科学院香港创新研究院）；Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；University of Chinese Academy of Sciences（中国科学院大学）；City University of Hong Kong Shenzhen Research Institute（香港城市大学深圳研究院）；University of Science and Technology of China（中国科学技术大学）

Song Lai, Zhe Zhao, Fei Zhu, Xi Lin, Qingfu Zhang, Gaofeng Meng

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出 ParetoCL 框架，将持续学习中的稳定性-可塑性权衡转化为多目标优化问题，通过偏好条件化模型学习 Pareto 最优解集，实现推理期间的动态适应，优于现有方法。

摘要：

持续学习旨在按顺序学习多个任务。其关键挑战在于平衡两个目标：保留旧任务知识（稳定性）与适应新任务（可塑性）。经验回放方法通过存储并重放过去数据与新数据，已成为缓解灾难性遗忘的广泛采用方法。然而，这些方法忽视了稳定性-可塑性权衡的动态特性，旨在寻找固定不变的平衡，导致训练和推理期间的适应次优。本文提出 Pareto Continual Learning (ParetoCL)，一个将持续学习中的稳定性-可塑性权衡重新表述为多目标优化 (MOO) 问题的新颖框架。ParetoCL 引入偏好条件化模型，高效学习代表不同权衡的 Pareto 最优解集，并实现推理期间的动态适应。从泛化视角看，ParetoCL 可视为一种目标增强方法，从稳定性和可塑性的不同目标组合中学习。跨多个数据集和设置的广泛实验表明，ParetoCL 优于 state-of-the-art 方法并适应多样的持续学习场景。

英文摘要：

Continual learning aims to learn multiple tasks sequentially. A key challenge in continual learning is balancing between two objectives: retaining knowledge from old tasks (stability) and adapting to new tasks (plasticity). Experience replay methods, which store and replay past data alongside new data, have become a widely adopted approach to mitigate catastrophic forgetting. However, these methods neglect the dynamic nature of the stability-plasticity trade-off and aim to find a fixed and unchanging balance, resulting in suboptimal adaptation during training and inference. In this paper, we propose Pareto Continual Learning (ParetoCL), a novel framework that reformulates the stability-plasticity trade-off in continual learning as a multi-objective optimization (MOO) problem. ParetoCL introduces a preference-conditioned model to efficiently learn a set of Pareto optimal solutions representing different trade-offs and enables dynamic adaptation during inference. From a generalization perspective, ParetoCL can be seen as an objective augmentation approach that learns from different objective combinations of stability and plasticity. Extensive experiments across multiple datasets and settings demonstrate that ParetoCL outperforms state-of-the-art methods and adapts to diverse continual learning scenarios.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。