研究人员推出ESI-Bench，这是一个旨在评估AI代理具身空间智能的新基准。该基准侧重于感知-行动闭环，代理通过主动探索环境来收集信息，而不是被动处理视觉数据。使用最先进的多模态大语言模型（MLLMs）进行的实验表明，与被动观察相比，主动探索显著提高了性能，尽管失败通常源于糟糕的行动选择而非薄弱的感知。该研究还强调了模型中存在的元认知差距，因为它们倾向于过早地做出结论，不像人类会根据矛盾的证据修正信念。 AI
影响 该基准有望推动能够进行更复杂现实世界交互和解决问题的AI代理的发展。
排序理由 该集群描述了一个用于评估AI代理空间智能的新基准，该基准在一篇学术论文中提出。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
Spatial intelligence unfolds through a perception-action loop: agents act to acquire observations, and reason about how observations vary as a function of action. Rather than passively processing what is seen, they actively uncover what is unseen - occluded structure, dynamics, c…
Spatial intelligence unfolds through a perception-action loop: agents act to acquire observations, and reason about how observations vary as a function of action. Rather than passively processing what is seen, they actively uncover what is unseen - occluded structure, dynamics, c…