ToolSelf: 通过工具驱动的涌现适应统一任务执行与自我重构
ToolSelf: Unifying Task Execution and Self-Reconfiguration via Tool-Driven Emergent Adaptation
AI总结：
提出ToolSelf框架，将配置更新抽象为标准化工具接口，统一任务执行与自我重构，并采用配置感知两阶段训练（CAT）实现涌现适应性，在多种基准测试中平均超越静态配置基线28.8分。
摘要：
基于LLM的智能体系统在复杂长时任务中表现出色，但仍受限于执行前固定的静态配置。这种刚性导致领域特定性能与跨任务泛化之间的权衡：强先验和紧凑工具空间有助于专业化但削弱迁移，而任务无关的工作流和广泛动作空间扩展覆盖但稀释指导。现有的执行前优化、规划者-工作者编排和配置修补未能解决这一矛盾，因为它们将适应与执行解耦，导致信息丢失、优化碎片化和信用分配模糊。我们提出ToolSelf，一种工具驱动的运行时自我重构范式，将配置更新抽象为标准化工具接口，并在一个策略的动作空间内统一执行和适应。执行代理可以根据任务进度和反馈动态更新子目标、策略、工具箱、上下文和上下文管理模式。我们进一步引入配置感知两阶段训练（CAT），结合拒绝采样微调和轨迹级KTO强化学习来内化自我重构。在多种基准测试中，零样本ToolSelf与任务专用代理相媲美；经过CAT训练后，ToolSelf平均比静态配置基线高出28.8分，为消除手动注入指导的涌现适应性开辟了道路。
英文摘要：
LLM-powered agentic systems excel at complex long-horizon tasks, but remain constrained by static configurations fixed before execution. Such rigidity forces a trade-off between domain-specific performance and cross-task generalization: strong priors and compact tool spaces aid specialization but weaken transfer, while task-agnostic workflows and broad action spaces expand coverage but dilute guidance. Existing pre-execution optimization, planner-worker orchestration, and configuration patching fall short of resolving this tension, as they decouple adaptation from execution, causing information loss, fragmented optimization, and ambiguous credit assignment. We propose ToolSelf, a tool-driven runtime self-reconfiguration paradigm that abstracts configuration updates as a standardized tool interface and unifies execution and adaptation within one policy's action space. The execution agent can dynamically update sub-goals, strategies, toolboxes, context, and context-management modes based on task progress and feedback. We further introduce Configuration-Aware Two-stage Training (CAT), which combines rejection sampling fine-tuning with trajectory-level KTO reinforcement learning to internalize self-reconfiguration. Across diverse benchmarks, zero-shot ToolSelf rivals task-specialized agents; after CAT training, ToolSelf gains 28.8 points over the static-configuration baseline on average, illuminating a path toward emergent adaptivity that obviates manually injected guidance. The code is available at https://github.com/lian-tian-mo-zun/ToolSelf.
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。
