小模型+流程设计，让诊断准确率超大模型
DeepLens Diagnosis Agent: Agentic Workflow Design Lets a Small Reasoning Model Compete with Frontier LLMs
- 用五阶段流程约束小模型推理，结合检索增强生成
- 诊断准确率达60.14%，比无流程时提升36个百分点
- 结果可追溯、可审计，适合医疗等高风险场景
医学诊断是多阶段过程：提取事实、查阅知识、生成鉴别分析并选出最佳诊断。前沿大模型虽强，但单次提示常导致脆弱推理。本文提出DeepLens诊断代理，基于小型医学推理模型（JSL Medical Small 7B v2）和检索增强生成（RAG），构建五阶段流程：结构化临床信息提取、受控检索、约束候选生成、显式证据交叉验证与可审计决策。在915例的DiagnosisArena基准上，该代理达到60.14%的顶1准确率，为小中型模型最高。同模型无流程时仅23.99%，流程设计带来+36分提升。尽管其标准医学基准得分88.2%，仍表明不确定性下的诊断需超越知识记忆。每例成本0.0072美元（A100，24K tokens），延迟24秒，比Claude Sonnet 4.5（0.0110美元）和Gemini 3.1 Pro（0.0128美元）便宜35-45%，且准确率分别高出9.70和9.17个百分点。流程设计亦能修正大模型错误，证明约束机制可超越参数量或成本。除整体性能外，流程生成结构化中间结果，支持各阶段可检查与错误定位，适用于对可追溯性、可复现性及可审计证据要求高的场景。
原文摘要 · Abstract (English)
Medical diagnosis is a multi-stage process: extract facts, consult knowledge, generate a differential analysis, and select the best diagnosis with explanations. Frontier LLMs are strong generalists, but single-shot prompting often yields brittle diagnostic reasoning. We present the DeepLens Diagnosis Agent, a five-stage harnessing pipeline (combining model capabilities with disciplined process constraints) centered on a small medical reasoning model (JSL Medical Small 7B v2) and retrieval-augmented generation (RAG). The pipeline enforces structured clinical extraction, disciplined retrieval, constrained candidate generation, explicit evidence triangulation, and an auditable final decision. On the 915-case DiagnosisArena benchmark, the agent achieved 60.14% top-1 diagnostic accuracy, the highest among small and medium-sized models. The same model without the agent workflow achieved 23.99%, a +36-point gain from workflow design alone, despite 88.2% on standard medical benchmarks, showing that diagnostic reasoning under uncertainty requires more than knowledge recall. The agent costs USD 0.0072 per case (24K tokens on A100) with 24-second latency, 35-45% cheaper than Claude Sonnet 4.5 (USD 0.0110) and Gemini 3.1 Pro (USD 0.0128) while outperforming them by +9.70pp and +9.17pp. Harnessing can also correct frontier model failures; workflow constraints can outweigh parameter count or API cost. Beyond aggregate accuracy, the pipeline produces structured intermediate artifacts that make each stage inspectable and support error localization. These properties support high-stakes settings where traceability, reproducibility, and auditable evidence matter alongside benchmark performance.
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。