arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2403.13249cs.LGcs.AIcs.CV
收藏

一个用于持续学习的统一且通用框架

A Unified and General Framework for Continual Learning

University of Maryland, College Park(马里兰大学帕克分校);JD Explore Academy(京东探索学院)

Zhenyi Wang, Yan Li, Li Shen, Heng Huang

AI总结:

本研究提出一个统一且通用的持续学习框架,将现有方法作为共享优化目标的特例包含在内,并引入受神经科学启发的、可无缝集成到现有CL方法中以以提升性能的刷新学习机制。

摘要:

持续学习(Continual Learning,简称CL)关注于从动态和变化的数据分布中学习,同时保留之前已获取的知识。。为了应对灾难性遗忘问题,人们开发了各种方法,包括基于正则化、基于贝叶斯和基于记忆回放的技术。然而,这些方法缺乏一个统一的框架和通用的术语来描述它们的方法。本研究旨在通过引入一个全面且具有统领性的框架来弥合这一差距,该框架包含并调和了现有的方法论。值得注意的是,这个新框架能够将现有的CL方法作为统一且通用优化目标的特例来包含在内。一个有趣的发现是:尽管这些方法来源各异,但它们共享相同的数学结构。这一观察揭示了这些看似不同的技术之间的兼容性,以及它们通过一个共享的底层优化目标相互关联。此外,所提出的通用框架引入了一种称为刷新学习(refresh learning)的创新概念,专门设计用于提升CL性能。这种新方法从神经科学中汲取灵感,在人脑中,遗忘通常会消除旧信息以改善对关键知识的保留,并促进新信息的获取。本质上,刷新学习首先首先初始学习当前数据,然后随后学习新数据,作为一个通用的插件,,无缝集成到现有的CL方法中,以增强和提升学习过程。在CL基准上的大量实验和理论分析证明了所提出的刷新学习的有效性。代码可在 https://github.com/Joey-wang123/CL-refresh-learning 获取

英文摘要:

Continual Learning (CL) focuses on learning from dynamic and changing data distributions while retaining previously acquired knowledge. Various methods have been developed to address the challenge of catastrophic forgetting, including regularization-based, Bayesian-based, and memory-replay-based techniques. However, these methods lack a unified framework and common terminology for describing their approaches. This research aims to bridge this gap by introducing a comprehensive and overarching framework that encompasses and reconciles these existing methodologies. Notably, this new framework is capable of encompassing established CL approaches as special instances within a unified and general optimization objective. An intriguing finding is that despite their diverse origins, these methods share common mathematical structures. This observation highlights the compatibility of these seemingly distinct techniques, revealing their interconnectedness through a shared underlying optimization objective. Moreover, the proposed general framework introduces an innovative concept called refresh learning, specifically designed to enhance the CL performance. This novel approach draws inspiration from neuroscience, where the human brain often sheds outdated information to improve the retention of crucial knowledge and facilitate the acquisition of new information. In essence, refresh learning operates by initially unlearning current data and subsequently relearning it. It serves as a versatile plug-in that seamlessly integrates with existing CL methods, offering an adaptable and effective enhancement to the learning process. Extensive experiments on CL benchmarks and theoretical analysis demonstrate the effectiveness of the proposed refresh learning. Code is available at \url{https://github.com/joey-wang123/CL-refresh-learning}.

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。

↑