EviSeek:以证据为中心的具身主动信息获取智能体框架
EviSeek: An Evidence-Centric Agentic Framework for Embodied Active Information Seeking
AI总结:
本文提出无需训练的EviSeek框架,将具身主动问答中交互的信息目的明确,在ESI-Bench数据集上,该框架采用Qwen3-VL-32B、Gemini 3.1 Pro时分别较单模型智能体提升7.1、3.7个百分点准确率,同时优化视觉支持加权准确率并减少平均token使用量。
摘要:
主动具身问答要求智能体通过视角变换与物理交互获取证据,核心挑战是避免无方向交互:采取合理行动却未明确其应揭示的、用于回答问题的证据。仅行动规划与累积历史会使交互的信息目的隐含。本文提出EviSeek,一种无需训练的框架,将问题衍生的证据需求明确为获取目标,将观测条件下的行动与选择性保留的证据关联至智能体寻求的信息。在3436个ESI-Bench问题上,采用Qwen3-VL-32B时,EviSeek较单模型智能体准确率提升7.1个百分点;采用Gemini 3.1 Pro时提升3.7个百分点,同时提升视觉支持加权准确率并减少平均token使用量。这些发现支持将交互的信息目的明确,围绕智能体回答问题所需学习的内容组织行动选择与证据保留。
英文摘要:
Active embodied question answering requires agents to acquire evidence through viewpoint changes and physical interaction. The central challenge is to avoid undirected interaction: taking plausible actions without specifying what evidence they should reveal for answering the question. Action plans and accumulated histories alone can leave the informational purpose of an interaction implicit. We introduce EviSeek, a training-free framework that makes question-derived evidence demands explicit acquisition objectives, linking observation-conditioned actions and selectively retained evidence to the information the agent seeks. On 3,436 ESI-Bench questions, EviSeek improves accuracy over the Single-model agent by 7.1 percentage points with Qwen3-VL-32B and 3.7 points with Gemini 3.1 Pro, while improving visual-support-weighted accuracy and reducing mean token usage. These findings support making the informational purpose of interaction explicit, with action selection and evidence retention organized around what the agent needs to learn to answer the question.
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。