cs.AINov 10, 2026

A computational framework for understanding embodied plans

Authors: Aalap D. Shah, Basil Saeed, Joshua B. Tenenbaum, Tobias Gerstenberg, Hyowon Gweon, Ilker Yildirim

Organizations: Department of Psychology, Yale University, New Haven, CT, USA. · Department of Statistics, Stanford University, Palo Alto, CA, USA. · Department of Brain & Cognitive Sciences, MIT, Cambridge, MA, USA. · Department of Psychology, Stanford University, Palo Alto, CA, USA. · 

Abstract

Reasoning about people's plans and goals based on their actions is fundamental to human social interaction and essential for cooperative, embodied artificial intelligence. Formalizing this capacity in naturalistic, embodied settings has been a longstanding challenge. Here, we present GenEA, a novel agent representation capable of reasoning over multi-step, long-horizon embodied plans. At the heart of GenEA lies a multilevel state-space that couples the coarse biomechanics of an articulated body and its interactions with nearby objects by extending the task-and-motion planning framework from robotics with simulation-based intuitive physics. By embedding this state-space in a generative model, GenEA yields a joint probability distribution over goals and embodied plans. Across several experiments spanning the qualitatively distinct domains of object manipulation and navigation, featuring diverse tasks and multiple behavioral measures, we show that GenEA, through simple transformations of this joint distribution, recapitulates how humans reason about embodied plans. Crucially, GenEA consistently outperforms strong alternative models, including plausible heuristics, standard task-and-motion planning, and a state-of-the-art vision-language model. GenEA thereby offers a new perspective on the architecture of human reasoning and social cognition, suggesting that embodied artificial intelligence systems seeking to seamlessly interact with humans should adopt GenEA-like generative models of action planning.

Figures & tables

Explore similar work

CardsList
  1. Generate-then-Rank: Learning Symbolic Goal Interpretation for Embodied Agents

    Nov 5, 2026Jisen Lei, Huizhi LiangEmbodied AISymbolic Reasoning in Language Models

  2. RoboAgent: Chaining Basic Capabilities for Embodied Task Planning

    Apr 9, 2026Peiran Xu, Jiaqi Zheng, Yadong MuRobot Task PlanningEmbodied Reasoning