📄 中文摘要
在使用大型语言模型（LLMs）开发自主智能体的过程中，诸如上下文窗口限制和会话碎片化等问题对知识的长期积累构成了重大障碍。该研究提出了一种“自我进化框架”，使智能体能够在不断变化的环境中持续学习和适应。通过递归知识晶化的过程，智能体能够有效整合和更新其知识库，从而提升决策能力和执行效率。这一框架为自主智能体的持续进化提供了理论基础和实践路径，具有广泛的应用前景。
递归知识晶化：持久自主智能体自我进化框架
📄 English Summary
Recursive Knowledge Crystallization: A Framework for Persistent Autonomous Agent Self-Evolution
The development of autonomous agents using Large Language Models (LLMs) faces significant barriers to long-term knowledge accumulation due to constraints like context window limits and session fragmentation. This study proposes a 'self-evolving framework' that enables agents to continuously learn and adapt in dynamic environments. Through a process of recursive knowledge crystallization, agents can effectively integrate and update their knowledge bases, enhancing their decision-making capabilities and execution efficiency. This framework provides a theoretical foundation and practical pathway for the persistent evolution of autonomous agents, with broad application prospects.