{"ID":2864539,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.23071","arxiv_id":"2509.23071","title":"From Evidence to Trajectory: Abductive Reasoning Path Synthesis for Training Retrieval-Augmented Generation Agents","abstract":"Retrieval-augmented generation agents development is hindered by the lack of process-level supervision to effectively guide agentic capabilities like task decomposition, retriever invocation, and stepwise decision-making. While reinforcement learning offers a potential solution, it suffers from sparse rewards and the limited reasoning capabilities of large language models (LLMs). Meanwhile, existing data synthesis methods only produce chain-of-thought rationales and fail to model environmental interactions. In this paper, we propose EviPath, an evidence-anchored reasoning path synthesis paradigm for RAG agent development. EviPath comprises: (i) Abductive Subtask Planning, which decomposes the problem into sub-questions and iteratively plans an optimal solution path based on the dependencies between them; (ii) Faithful Sub-question Answering, which uses supporting evidence to construct a proxy environment to generate reasoning thoughts and answers for each sub-question; and (iii) Conversational Fine-Tuning, which formats the complete agent-environment interaction trajectory into a dialogue format suitable for Supervised Fine-Tuning. EviPath allows LLMs to learn complex reasoning and tool-use capabilities directly from synthesized data. Extensive experiments on widely-used question-answering benchmarks show that an 8B parameter model trained with EviPath-synthesized data significantly and consistently outperforms state-of-the-art baselines with a double-digit absolute EM gain of 14.7% in open-domain question answering.","short_abstract":"Retrieval-augmented generation agents development is hindered by the lack of process-level supervision to effectively guide agentic capabilities like task decomposition, retriever invocation, and stepwise decision-making. While reinforcement learning offers a potential solution, it suffers from sparse rewards and the l...","url_abs":"https://arxiv.org/abs/2509.23071","url_pdf":"https://arxiv.org/pdf/2509.23071v1","authors":"[\"Muzhi Li\",\"Jinhu Qi\",\"Yihong Wu\",\"Minghao Zhao\",\"Liheng Ma\",\"Yifan Li\",\"Xinyu Wang\",\"Yingxue Zhang\",\"Ho-fung Leung\",\"Irwin King\"]","published":"2025-09-27T02:53:09Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\"]","methods":"[\"RAG\",\"Reinforcement Learning\",\"Large Language Model\",\"Language Model\"]","has_code":false}