CHEEM：通过重用、新信息、适应和跳过进行持续学习——一种分层探索-利用方法
CHEEM: Continual Learning by Reuse, New, Adapt and Skip -- A Hierarchical Exploration-Exploitation Approach
AI总结：
本文提出CHEEM框架，通过分层探索-利用方法解决持续学习中的稳定性与可塑性矛盾，实现高效神经网络架构搜索，优于现有方法，接近全微调上限。
摘要：
为了有效管理现实动态环境的复杂性，持续学习必须从不同性质的任务流中逐步获取、更新和积累知识，而不陷入灾难性遗忘。尽管人类认知具备此能力，但现代深度学习系统仍面临显著挑战。核心挑战在于平衡利用先前知识、整合新信息以及根据任务复杂性和协同效应分配模型容量。本文提出了一种无需示例的类增量持续学习（ExfCCL）框架，通过分层探索-利用（HEE）方法解决这些问题。方法的核心是HEE引导的高效神经网络架构搜索（HEE-NAS），通过四个基本操作——重用、新信息、适应和跳过——实现学习-适应的骨干网络，作为内部记忆动态更新所选组件。为了解决ExfCCL中的任务ID推断问题，我们利用先前研究中提出的外部记忆任务质心。我们的方法称为CHEEM（持续分层探索-利用记忆）。CHEEM在挑战性的MTIL和VDD基准上使用Tiny和Base Vision Transformers以及提出的综合Figure-of-Merit（FoM）度量进行评估。它显著优于最先进的提示基于持续学习方法，接近全微调上限。此外，它以语义有意义的方式为每个任务学习适应的模型结构。我们的代码可在https://github.com/savadikarc/cheem获取。
英文摘要：
To effectively manage the complexities of real-world dynamic environments, continual learning must incrementally acquire, update, and accumulate knowledge from a stream of tasks of different nature without suffering from catastrophic forgetting of prior knowledge. While this capability is innate to human cognition, it remains a significant challenge for modern deep learning systems. At the heart of this challenge lies the stability-plasticity dilemma: the need to balance leveraging prior knowledge, integrating novel information, and allocating model capacity adaptively based on task complexity and synergy. In this paper, we propose a novel exemplar-free class-incremental continual learning (ExfCCL) framework that addresses these issues through a Hierarchical Exploration-Exploitation (HEE) approach. The core of our method is a HEE-guided efficient neural architecture search (HEE-NAS) that enables a learning-to-adapt backbone via four primitive operations - reuse, new, adapt, and skip - thereby serving as an internal memory that dynamically updates selected components across streaming tasks. To address the task ID inference problem in ExfCCL, we exploit an external memory of task centroids proposed in the prior art. We term our method CHEEM (Continual Hierarchical-Exploration-Exploitation Memory). CHEEM is evaluated on the challenging MTIL and VDD benchmarks using both Tiny and Base Vision Transformers and a proposed holistic Figure-of-Merit (FoM) metric. It significantly outperforms state-of-the-art prompting-based continual learning methods, closely approaching full fine-tuning upper bounds. Furthermore, it learns adaptive model structures tailored to individual tasks in a semantically meaningful way. Our code is available at https://github.com/savadikarc/cheem .
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。
