智能体KB：利用跨领域经验实现自主问题解决

摘要

随着语言代理处理日益复杂的任务，它们在有效纠错和跨领域经验复用方面面临挑战。我们引入了Agent KB，一个层次化的经验框架，通过新颖的“推理-检索-精炼”流程实现复杂的代理问题解决。Agent KB解决了一个核心限制：传统上代理无法从彼此的经验中学习。通过捕捉高层策略和详细执行日志，Agent KB创建了一个共享知识库，促进了跨代理的知识转移。在GAIA基准测试中，Agent KB将成功率提升了高达16.28个百分点。在最具挑战性的任务上，Claude-3的成功率从38.46%提升至57.69%，而GPT-4在中等难度任务上从53.49%提升至73.26%。在SWE-bench代码修复任务中，Agent KB使Claude-3的成功率从41.33%提升至53.33%。我们的结果表明，Agent KB提供了一个模块化、框架无关的基础设施，使代理能够从过往经验中学习，并将成功策略泛化至新任务。

English

As language agents tackle increasingly complex tasks, they struggle with effective error correction and experience reuse across domains. We introduce Agent KB, a hierarchical experience framework that enables complex agentic problem solving via a novel Reason-Retrieve-Refine pipeline. Agent KB addresses a core limitation: agents traditionally cannot learn from each other's experiences. By capturing both high-level strategies and detailed execution logs, Agent KB creates a shared knowledge base that enables cross-agent knowledge transfer. Evaluated on the GAIA benchmark, Agent KB improves success rates by up to 16.28 percentage points. On the most challenging tasks, Claude-3 improves from 38.46% to 57.69%, while GPT-4 improves from 53.49% to 73.26% on intermediate tasks. On SWE-bench code repair, Agent KB enables Claude-3 to improve from 41.33% to 53.33%. Our results suggest that Agent KB provides a modular, framework-agnostic infrastructure for enabling agents to learn from past experiences and generalize successful strategies to new tasks.

智能体KB：利用跨领域经验实现自主问题解决

Agent KB: Leveraging Cross-Domain Experience for Agentic Problem Solving

摘要

Support