arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2604.07595cs.AIcs.CL
收藏

推理图:通过以证据为中心的反馈实现自我改进的确定性RAG

Reasoning Graphs: Self-Improving, Deterministic RAG through Evidence-Centric Feedback

Matthew Penaroza

AI总结:

本文提出推理图,通过证据中心反馈提升RAG的准确性和确定性,实验显示在50%以上的证据覆盖下,错误率降低47%,在多跳问题中准确率提升11.0个百分点,实现高重用场景下的性能优势。

摘要:

语言模型代理在每次查询中从头开始推理,丢弃其思维链,导致准确性低和高方差,因为相同查询类型可能成功或失败不可预测。我们引入推理图,一种图结构,每条证据链的思维过程以结构化的边连接到评估的证据项。与之前基于查询相似性的记忆机制不同,推理图支持以证据为中心的反馈:给定新的候选集,系统遍历所有先前运行中每个证据项的输入评估边,揭示该特定项之前如何被判断。我们进一步引入检索图,一种互补结构,将管道规划器输入以收紧后续运行的候选 funnel。两者共同形成自我改进的反馈循环:准确性系统性提高,判断层面的方差崩溃。这不需要重新训练;基础模型保持冻结,所有收益来自上下文工程通过图遍历。我们在MuSiQue和HotpotQA上进行评估,使用顺序聚类协议,高重用部署模拟和确定性实验。在50%+证据配置文件覆盖下,我们的系统与vanilla RAG相比,相同问题的错误率降低47%(受控剂量反应,p < 0.0001)。在4跳问题中,准确率提高11.0pp(p=0.0001)。在高重用场景中,系统实现帕累托支配:最高准确率,47%更低成本,46%更低延迟。证据配置文件通过7-8个百分点提高判断一致性(p=0.007,Wilcoxon);整个系统驱动所有11个硬探针在温度0和0.5时达到完美一致性(p=0.004)

英文摘要:

Language model agents reason from scratch on every query, discarding their chain of thought after each run. This produces lower accuracy and high variance, as the same query type can succeed or fail unpredictably. We introduce reasoning graphs, a graph structure that persists per-evidence chain of thought as structured edges connected to the evidence items they evaluate. Unlike prior memory mechanisms that retrieve distilled strategies by query similarity, reasoning graphs enable evidence-centric feedback: given a new candidate set, the system traverses all incoming evaluation edges for each evidence item across all prior runs, surfacing how that specific item has been judged before. We further introduce retrieval graphs, a complementary structure that feeds a pipeline planner to tighten the candidate funnel over successive runs. Together, both graphs form a self-improving feedback loop: accuracy improves systematically and verdict-level variance collapses. This requires no retraining; the base model remains frozen and all gains come from context engineering via graph traversal. We evaluate on MuSiQue and HotpotQA using a sequential cluster protocol, a high-reuse deployment simulation, and a determinism experiment. At 50%+ evidence profile coverage, our system reduces errors by 47% compared to vanilla RAG on the same questions (controlled dose-response, p < 0.0001). On 4-hop questions, accuracy improves by +11.0pp (p=0.0001). In high-reuse settings, the system achieves Pareto dominance: highest accuracy, 47% lower cost, and 46% lower latency. Evidence profiles improve verdict consistency by 7-8 percentage points (p=0.007, Wilcoxon); the full system drives all 11 hard probes to perfect consistency at both temperature 0 and 0.5 (p=0.004).

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。

↑