选择合适的专家：基于神经过程的注意力机制用于选择任务专用模型作为智能医疗系统工具
Picking the Right Specialist: Attentive Neural Process-based Selection of Task-Specialized Models as Tools for Agentic Healthcare Systems
AI总结：
本文提出ToolSelect，一种基于神经过程和注意力机制的模型选择方法，用于智能医疗系统中选择任务专用模型，通过实验展示其在不同任务上的优越性能。
摘要：
任务专用模型是智能医疗系统的核心，使代理能够回答涉及疾病诊断、定位和报告生成等任务的临床问题。然而，对于给定的任务，很少有单一“最佳”模型存在。实际上，每个任务都更适合多个竞争的专家模型，不同的模型在不同的数据样本上表现各异。因此，对于任何给定的查询，代理必须可靠地从异构的工具候选池中选择正确的专家模型。为此，我们引入了ToolSelect，它通过最小化在采样专家工具候选上的人口风险，利用一致的替代任务条件选择损失来适应性学习模型选择。具体来说，我们提出了一种基于注意力机制的神经过程选择器，该选择器根据查询和每种模型的行为摘要来选择专家模型。受缺乏任何已建立测试平台的启发，我们首次引入了一个配备多样任务专用模型（17个疾病检测、19个报告生成、6个视觉定位和13个VQA）的智能胸片环境，并开发了ToolSelectBench，一个包含1448个查询的基准测试。我们的结果表明，ToolSelect在四个不同的任务家族中一致地优于10种最先进的方法。
英文摘要：
Task-specialized models form the backbone of agentic healthcare systems, enabling the agents to answer clinical queries across tasks such as disease diagnosis, localization, and report generation. Yet, for a given task, a single "best" model rarely exists. In practice, each task is better served by multiple competing specialist models where different models excel on different data samples. As a result, for any given query, agents must reliably select the right specialist model from a heterogeneous pool of tool candidates. To this end, we introduce ToolSelect, which adaptively learns model selection for tools by minimizing a population risk over sampled specialist tool candidates using a consistent surrogate of the task-conditional selection loss. Concretely, we propose an Attentive Neural Process-based selector conditioned on the query and per-model behavioral summaries to choose among the specialist models. Motivated by the absence of any established testbed, we, for the first time, introduce an agentic Chest X-ray environment equipped with a diverse suite of task-specialized models (17 disease detection, 19 report generation, 6 visual grounding, and 13 VQA) and develop ToolSelectBench, a benchmark of 1448 queries. Our results demonstrate that ToolSelect consistently outperforms 10 SOTA methods across four different task families.
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。
