AI 中文总结
针对LLM智能体依赖静态工具空间难以适配开放世界科学工作流的问题，提出SciToolAgent-Evo本体感知自进化智能体，结合进化记忆与本体化工具图，利用LinUCB平衡探索利用，推出OpenSciToolBench基准并取得最优性能。
AI 中文摘要
大型语言模型（LLM）智能体已越来越多地应用于科学研究中，用于组织和调用专用计算工具。然而，它们依赖具有静态语义的预定义工具空间，这限制了其在开放世界科学工作流程中的适用性，这类流程中工具需求、能力和边界会动态变化。为此，我们提出SciToolAgent-Evo，一种面向开放世界科学工具获取的本体感知自进化智能体。它由技能、经验的进化记忆以及本体化工具图驱动，在积累过程中从对比轨迹中提炼可泛化知识；推理时，它会制定主动请求并利用基于LinUCB的多臂赌博机门来动态平衡探索与利用。一旦获取新工具，其科学本体会在线完成，以无缝集成到已知图中。此外，我们推出OpenSciToolBench，一个包含四个难度级别共900个真实任务的基准。大量评估表明，SciToolAgent-Evo达到了最先进的性能，验证了其鲁棒性和泛化能力。
英文摘要
Large language model (LLM) agents have been increasingly adopted in scientific research for organizing and invoking specialized computational tools. However, their reliance on predefined tool spaces with static semantics limits their applicability to open-world scientific workflows, where tool requirements, capabilities, and boundaries evolve dynamically. To this end, we propose SciToolAgent-Evo, an ontology-aware self-evolving agent for open-world scientific tool acquisition. Driven by an evolving memory of skills, experiences, and an ontologized tool graph, it distills generalizable knowledge from contrastive trajectories during accumulation, whereas during inference, it formulates active requests and utilizes a LinUCB-based bandit gate to dynamically balance exploration and exploitation. Once a novel tool is acquired, its scientific ontology is completed online for seamless integration into the known graph. Moreover, we introduce OpenSciToolBench, a benchmark containing 900 realistic tasks across four difficulty levels. Extensive evaluations show that SciToolAgent-Evo achieves state-of-the-art performance, validating its robustness and generalization.
Comments19 pages, 4 figures, under review
