用智能代理工具增强大模型推理，解决复杂问题。
Agentic Reasoning: A Streamlined Framework for Enhancing LLM Reasoning with Agentic Tools
- 引入动态调用搜索、代码执行和结构化记忆的智能代理框架。
- 在DeepSeek-R1上达到公开模型新最优，接近闭源领先模型表现。
- 创新思维图代理保持长链推理连贯性，适合复杂研究任务。
我们提出Agentic Reasoning框架，通过集成外部工具型智能体来增强大语言模型（LLM）的推理能力。该框架动态调用网络搜索、代码执行和结构化记忆，以应对需要深入调研的复杂问题。核心创新是思维图代理（Mind-Map agent），它构建结构化知识图谱，存储推理上下文并追踪逻辑关系，确保在大量工具使用下的长链推理连贯性。此外，我们对网络搜索代理进行了全面探索，提出一种高效搜索机制，超越所有先前方法。在DeepSeek-R1上部署后，该方法在公开模型中达到新的最先进水平，性能接近OpenAI Deep Research这一领域领先的专有模型。大量消融实验验证了智能体工具的最佳选择，并确认思维图与网络搜索代理在提升推理能力上的有效性。代码已开源：https://github.com/theworldofagents/Agentic-Reasoning。
原文摘要 · Abstract (English)
We introduce Agentic Reasoning, a framework that enhances large language model (LLM) reasoning by integrating external tool-using agents. Agentic Reasoning dynamically leverages web search, code execution, and structured memory to address complex problems requiring deep research. A key innovation in our framework is the Mind-Map agent, which constructs a structured knowledge graph to store reasoning context and track logical relationships, ensuring coherence in long reasoning chains with extensive tool usage. Additionally, we conduct a comprehensive exploration of the Web-Search agent, leading to a highly effective search mechanism that surpasses all prior approaches. When deployed on DeepSeek-R1, our method achieves a new state-of-the-art (SOTA) among public models and delivers performance comparable to OpenAI Deep Research, the leading proprietary model in this domain. Extensive ablation studies validate the optimal selection of agentic tools and confirm the effectiveness of our Mind-Map and Web-Search agents in enhancing LLM reasoning. The code is at: https://github.com/theworldofagents/Agentic-Reasoning
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。