让AI动态决定如何构建知识图谱上下文，更快更省地答对复杂问题。
CLAUSE: Agentic Neuro-Symbolic Knowledge Graph Reasoning via Dynamic Learnable Context Engineering
- 用三个智能体协作，按需扩展知识图谱路径，避免盲目搜索。
- 在相同成本下，准确率提升39.3%，延迟降18.6%，图增长少40.9%。
- 适合需要快速、可控、可追溯推理的部署场景，如客服与医疗问答。
知识图谱为多跳问答提供结构化上下文，但实际系统需在答案准确性、严格延迟与成本目标之间权衡，同时保留溯源信息。静态的k跳扩展和“思考更久”提示常导致过度检索、上下文膨胀及运行时间不可预测。我们提出CLAUSE，一种基于动态可学习上下文工程的代理型神经符号框架，将上下文构建视为知识图谱上的序列决策过程，动态决定扩展内容、路径选择或回溯、保留证据时机及停止条件。延迟（交互步数）和提示成本（选中词元数）作为用户指定预算或价格，支持每查询自适应权衡准确性、延迟与成本，无需重新训练。CLAUSE采用提出的拉格朗日约束多智能体近端策略优化（LC-MAPPO）算法，协调子图架构师、路径导航者与上下文整理者三个智能体，联合优化子图构建、推理路径发现与证据选择，受每查询资源预算（边修改、交互步数、选中词元）约束。在HotpotQA、MetaQA和FactKG上，CLAUSE在同等或更低词元预算下，提升EM@1准确率，减少子图增长与端到端延迟。在MetaQA-2-hop上，相比最强RAG基线GraphRAG，CLAUSE实现+39.3 EM@1，延迟降低18.6%，边增长减少40.9%。生成的上下文紧凑、可溯源，且在部署约束下表现稳定可预测。
原文摘要 · Abstract (English)
Knowledge graphs provide structured context for multi-hop question answering, but deployed systems must balance answer accuracy with strict latency and cost targets while preserving provenance. Static k-hop expansions and "think-longer" prompting often over-retrieve, inflate context, and yield unpredictable runtime. We introduce CLAUSE, an agentic three-agent neuro-symbolic framework that treats context construction as a sequential decision process over knowledge graphs, deciding what to expand, which paths to follow or backtrack, what evidence to keep, and when to stop. Latency (interaction steps) and prompt cost (selected tokens) are exposed as user-specified budgets or prices, allowing per-query adaptation to trade-offs among accuracy, latency, and cost without retraining. CLAUSE employs the proposed Lagrangian-Constrained Multi-Agent Proximal Policy Optimization (LC-MAPPO) algorithm to coordinate three agents: Subgraph Architect, Path Navigator, and Context Curator, so that subgraph construction, reasoning-path discovery, and evidence selection are jointly optimized under per-query resource budgets on edge edits, interaction steps, and selected tokens. Across HotpotQA, MetaQA, and FactKG, CLAUSE yields higher EM@1 while reducing subgraph growth and end-to-end latency at equal or lower token budgets. On MetaQA-2-hop, relative to the strongest RAG baseline (GraphRAG), CLAUSE achieves +39.3 EM@1 with 18.6% lower latency and 40.9% lower edge growth. The resulting contexts are compact, provenance-preserving, and deliver predictable performance under deployment constraints.
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。