arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2606.18947cs.AIcs.CLcs.IRcs.MA

将搜索与推理解耦:面向LLM Agent的供应商无关的接地架构

Decoupling Search from Reasoning: A Vendor-Agnostic Grounding Architecture for LLM Agents

DoorDash, Inc.(DoorDash公司)

Emmanuel Aboah Boateng, Kyle MacDonald, Amardeep Kumar, Siddharth Kodwani, Sudeep Das

AI总结:

提出解耦搜索接地(DSG)架构,将搜索接地从推理模型中分离,通过MCP兼容网关实现供应商路由、缓存等控制,在降低成本和延迟的同时保持或提升准确性。

摘要:

生产级LLM Agent越来越依赖实时搜索,但原生搜索接地将检索策略、供应商选择、证据注入、成本、延迟和生成行为捆绑在单一模型-供应商边界内。这种耦合使得接地难以检查、调优、重用或移植,并可能触发搜索诱导的冗长,破坏严格的输出合约。我们提出解耦搜索接地(DSG),一种供应商无关的边界,通过MCP兼容网关将接地移出推理模型,将供应商路由、源感知上下文渲染、配置的回退、检索深度控制以及精确和语义缓存作为一级控制暴露。在SimpleQA、FreshQA和HotpotQA上的五个前沿模型上,原生搜索在时效性敏感的FreshQA上领先,但DSG在控制重要时展现出更强的前沿:在SimpleQA上,它以91%更低的搜索成本接近原生准确率(86.1%对87.7%),保持简洁答案合约,并以68%更低的延迟达到99.4%的热缓存命中率。作为大规模Agent工作负载的共享生产接地层部署,DSG在电商查询理解(QIU)工作负载上匹配或略超原生搜索准确率,同时将搜索成本降低超过98%。实时接地最好被视为可优化的接口边界,而非固定的模型特性。

英文摘要:

Production LLM agents increasingly depend on real-time search, yet native search grounding bundles retrieval policy, provider choice, evidence injection, cost, latency, and generation behavior behind a single model-provider boundary. This coupling makes grounding hard to inspect, tune, reuse, or port, and can trigger Search-Induced Verbosity that breaks strict output contracts. We present Decoupled Search Grounding (DSG), a vendor-agnostic boundary that moves grounding outside the reasoning model through an MCP-compatible gateway, exposing provider routing, source-aware context rendering, configured fallback, retrieval-depth control, and exact plus semantic caching as first-class controls. Across five frontier models on SimpleQA, FreshQA, and HotpotQA, native search leads on recency-sensitive FreshQA, but DSG exposes a stronger frontier when control matters: on SimpleQA it nearly matches native accuracy (86.1% vs. 87.7%) at 91% lower search cost, preserves concise answer contracts, and reaches a 99.4% warm-cache hit rate with 68% lower latency. Deployed as a shared production grounding layer for large-scale agentic workloads with interchangeable models, DSG matches or slightly exceeds native-search accuracy on an e-commerce query-understanding (QIU) workload while cutting search cost by over 98%. Real-time grounding is best treated as an optimizable interface boundary, not a fixed model feature.

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。

↑