arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.12658cs.CLcs.AI
收藏

在大型语言模型中进行持续学习:方法、挑战与机遇

Beyond Static Models: An Evolving Framework for Continual Learning in Large Language Models across Training Stages

Research Center for Scientific Data Hub, Zhejiang Lab(科学数据枢纽研究中心,浙江实验室)

Hongyang Chen, Zhongwu Sun, Hongfei Ye, Kunchi Li, Xuemin Lin

AI总结:

本文综述了针对大型语言模型的持续学习方法,分析了持续预训练、微调和对齐的核心阶段,探讨了传统方法在持续学习中的适应性与改进,并指出在跨任务和时间尺度的知识整合中仍存在挑战。

摘要:

持续学习(CL)已成为一种关键范式,使大型语言模型(LLMs)能够动态适应不断演变的知识和连续任务,同时缓解灾难性遗忘——这是现代LLMs静态预训练范式的关键限制。本文综述了针对LLMs的持续学习方法,围绕三个核心训练阶段:持续预训练、持续微调和持续对齐进行全面概述。除了传统的回放、正则化和架构基于方法之外,我们进一步将每个类别按其独特的遗忘缓解机制细分,并对传统持续学习方法在LLMs中的适应性和关键改进进行了严格比较分析。在此过程中,我们明确突出了持续学习与传统机器学习之间的核心区别,特别是在规模、参数效率和涌现能力方面。我们的分析涵盖了包括遗忘率和知识转移效率在内的关键评估指标,以及新兴基准以评估持续学习性能。本文揭示了尽管当前方法在特定领域表现出色,但在实现跨多样任务和时间尺度的无缝知识整合方面仍存在根本挑战。本文系统综述为LLM适应领域的知识库做出了贡献,为研究人员和实践者提供了理解当前成就和未来语言模型终身学习机会的结构化框架。

英文摘要:

Continual learning (CL) has emerged as a pivotal paradigm to enable large language models (LLMs) to dynamically adapt to evolving knowledge and sequential tasks while mitigating catastrophic forgetting, a critical limitation of the static pre-training paradigm inherent to modern LLMs. This paper presents a comprehensive overview of CL methodologies tailored for LLMs, structured around three core training stages: continual pre-training, continual fine-tuning, and continual alignment. Beyond the canonical taxonomy of rehearsal-, regularization-, and architecture-based methods, we further subdivide each category by its distinct forgetting mitigation mechanisms and conduct a rigorous comparative analysis of the adaptability and critical improvements of traditional CL methods for LLMs. In doing so, we explicitly highlight core distinctions between LLM CL and traditional machine learning, particularly with respect to scale, parameter efficiency, and emergent capabilities. Our analysis covers essential evaluation metrics, including forgetting rates and knowledge transfer efficiency, along with emerging benchmarks for assessing CL performance. While current methods show promising results, fundamental challenges persist in achieving seamless knowledge integration, and we identify key open problems and future directions for the field.

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。

↑