通过证据驱动的反馈图，让RAG系统越用越准、越用越快。
ROZA Graphs: Self-Improving Near-Deterministic RAG through Evidence-Centric Feedback
- 构建持续留存证据推理路径的图结构，实现基于历史判断的精准反馈。
- 在相同问题上达到10.6个百分点的准确率提升，47%错误率下降。
- 适合追求高稳定性和低延迟的生产级RAG部署场景。
语言模型代理每次查询都从零开始推理，推理过程不保留，导致准确率低且结果波动大。本文提出推理图，将每个证据的思考路径以结构化边的形式持久化。与以往依赖查询相似性召回策略的记忆机制不同，推理图支持证据中心的反馈：对每个候选项，系统遍历其过往所有评估边，回溯该条目此前被如何评判。进一步引入检索图，由规划器在多轮中剔除始终被拒绝的候选项。二者共同构成ROZA图：一个自我优化的反馈循环，其中准确率提升与黄金段落复用率正相关（推理图），效率提升与候选池重叠度正相关（检索图）。基础模型保持冻结，所有收益来自通过图遍历实现的上下文工程。在MuSiQue、HotpotQA及高复用部署子集上评估，发现四点显著成果：(1) 剂量反应关系：准确率随证据覆盖度单调上升，在50%以上覆盖率下相比基线提升+10.6个百分点（误差降低47%，p<0.0001；每题斯皮尔曼相关ρ=+0.144，p<10⁻⁶，n=1,100）；(2) 多跳推理扩展：4跳准确率提升+11.0个百分点（p=0.0001）；(3) 跨聚类预测：聚类级增益可由黄金段落复用密度预测（r=0.604，p=0.001，n=26个聚类）；(4) 高复用帕累托优势：在最高或并列最高准确率的同时，成本降低46%，延迟降低46%。固定20段落上下文下，跨重复运行的每段决策一致性提升8至13个百分点；当检索图同时进行剪枝时，提升达12至21个百分点（所有p<0.005，N=73配对探针，K=10次运行，两种模型族，三种温度）。
原文摘要 · Abstract (English)
Language model agents reason from scratch on every query, discarding their chain of thought after each run. The result is lower accuracy and high run-to-run variance. We introduce reasoning graphs, which persist the per-evidence chain of thought as structured edges. Unlike prior memory that retrieves distilled strategies by query similarity, reasoning graphs enable evidence-centric feedback: for every candidate item, the system traverses all incoming evaluation edges across prior runs to surface how that specific item has been judged before. We further introduce retrieval graphs, which feed a planner that prunes consistently-rejected candidates over successive runs. Together they form a ROZA graph: a self-improving feedback loop in which accuracy gains scale with gold-passage reuse (reasoning graph) and efficiency gains scale with candidate-pool overlap (retrieval graph). The base model remains frozen; all gains come from context engineering via graph traversal. We evaluate on MuSiQue and HotpotQA, plus a high-reuse deployment subset. Four findings stand out. (1) Dose-response: accuracy improves monotonically with evidence-profile coverage, reaching +10.6pp over Vanilla RAG at 50%+ coverage on the same questions (47% error reduction, $p<0.0001$; per-question Spearman $ρ=+0.144$, $p<10^{-6}$, $n=1{,}100$). (2) Multi-hop scaling: 4-hop accuracy improves by +11.0pp ($p=0.0001$). (3) Cross-cluster prediction: the cluster-level gain is predicted by gold-passage reuse density ($r=0.604$, $p=0.001$, $n=26$ clusters). (4) High-reuse Pareto dominance: highest or tied-for-highest accuracy alongside 46% lower cost and 46% lower latency. Per-passage decision consistency across repeated runs ($N=73$ paired probes, $K=10$ runs each, two model families, three temperatures) rises by +8 to +13pp on a fixed 20-passage context and by +12 to +21pp when the retrieval graph also prunes (all $p<0.005$).
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。