Large language model agents are often deployed in environments where new knowledge, tools, and workflows keep arriving. Such an agent is therefore expected to adapt continually, learning new tasks as they arrive without forgetting earlier ones. This resembles the continual learning problem, but at a pace that conventional gradient updates cannot sustain. Nonparametric self-evolving agents are a natural alternative: they adapt by rewriting their own harness, such as prompts, memory, and skills, from task feedback while the model weights stay frozen. We study whether such agents are effective continual learners across four drifting task streams that shift in topic, workflow, or domain (built from BIRD, BFCL, and ScienceWorld). Evaluating three representative methods spanning experience retrieval and self-modification (ReasoningBank, GEPA, and Meta-Harness), we find that although self-evolution improves aggregate performance on most streams, higher plasticity consistently comes at the cost of stability. Specifically, self-modifying methods apply global edits to a shared prompt or workspace, interfering with unseen tasks and overwriting earlier knowledge as the stream shifts. Conversely, similarity-based memory retrieval localizes updates across distant domains, but still retrieves foreign memories on unseen tasks and suffers from retrieval collision when similar queries require conflicting workflows. These findings show that current self-evolving agents remain incomplete continual learners.
SLCA: Slow Learner with Classifier Alignment for Continual Learning on a Pre-trained Model
University of Technology Sydney(悉尼科技大学);Tsinghua University(清华大学);Beihang University(北京航空航天大学);Zhongguancun Laboratory(中关村实验室);Beijing Jiaotong University(北京交通大学)
Learning to Continually Learn Rapidly from Few and Noisy Data
The Australian National University(澳大利亚国立大学);Monash University(莫纳什大学);CSIRO(澳大利亚联邦科学与工业研究组织);University of Turku(图尔库大学)
神经网络遭受灾难性遗忘,并且在数据分布中没有保证平稳性的情况下无法顺序地学习新任务。持续学习可以通过重放来实现--在学习新任务的同时同时训练外部存储的旧数据。然而,当每个过去的任务分配较少的内存时,重放变得不那么有效。为了克服这个困难,我们补充了重放机制与元学习快速知识获取。通过采用元学习者,即\textit{learns a learning rate per parameter per past task},我们发现,基础学习者产生强大的结果时,较少的内存可用。此外,我们的方法继承了持续学习的几个元学习优势:它表现出很强的鲁棒性,在噪声的存在下不断学习,并在较少的更新中获得更高的准确性。
CMKL: Modality-Aware Continual Learning for Evolving Biomedical Knowledge Graphs
Technology, Innovation, Entrepreneurship Department(技术、创新与创业部门);King Abdullah University of Science and Technology(卡塔尔科技大学);School of Computing and Information Systems(计算与信息系统学院);The University of Melbourne(墨尔本大学);College of Computer Science and Technology(计算机科学与技术学院);Jilin University(吉林大学);School of Computing Technologies(计算技术学院);RMIT University(皇家墨尔本理工大学);College of Computing and Data Science(计算与数据科学学院);Nanyang Technological University(南洋理工大学)
ContiGuard: A Framework for Continual Toxicity Detection Against Evolving Evasive Perturbations
School of Computer Science(计算机科学学院);Wuhan University(武汉大学);Zhongguancun Academy(中关村学院);Key Laboratory of Computing Power Network and Information Security, Ministry of Education, Shandong Computer Science Center (National Supercomputer Center in Jinan)(计算能力网络与信息安全重点实验室,教育部,山东计算机科学中心(济南国家超级计算机中心));Qilu University of Technology (Shandong Academy of Sciences)(齐鲁工业大学(山东科学院));Shandong Provincial Key Laboratory of Computing Power Internet and Service Computing(山东省计算能力互联网与服务计算重点实验室);Shandong Fundamental Research Center for Computer Science(山东省计算机科学基础研究中心)
When the Critic Falls Behind: Co-Evolving Rubrics and Harnesses for Continual Agent Self-Improvement
Research Center for Social Computing and Interactive Robotics, Harbin Institute of Technology(哈尔滨工业大学社会计算与交互机器人研究中心);Department of Computer Science and Technology, Institute for AI, Tsinghua University(清华大学计算机科学与技术系、人工智能研究院)
CoMemNet: A Continual Memory Network with Drift-Aware Sampling for Traffic Prediction
College of Computer Science, Shanghai Jiao Tong University(上海交通大学计算机科学学院);College of Computer Science and Technology, Zhejiang University of Technology(浙江工业大学计算机科学与技术学院);School of Automation, Nanjing University of Science and Technology(南京理工大学自动化学院)
From a Prompt to Repertoires: Evolving Functional REpertoires Enable LLM Continual Learning
The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳));Shanghai AI Laboratory(上海人工智能实验室);University of Science and Technology of China(中国科学技术大学);New Jersey Institute of Technology(新泽西理工学院)
Rethinking Continual Experience Internalization for Self-Evolving LLM Agents
Gaoling School of Artificial Intelligence, Renmin University of China(中国人民大学人工智能学院 Gallagher 学院);School of Software, Beihang University(北航软件学院);Meituan(美团)