ENACT: Evaluating Embodied Cognition with World Modeling of Egocentric Interaction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Qineng, Huang, Wenlong, Zhou, Yu, Yin, Hang, Bao, Tianwei, Lyu, Jianwen, Liu, Weiyu, Zhang, Ruohan, Wu, Jiajun, Fei-Fei, Li, Li, Manling
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911287422746624
author Wang, Qineng
Huang, Wenlong
Zhou, Yu
Yin, Hang
Bao, Tianwei
Lyu, Jianwen
Liu, Weiyu
Zhang, Ruohan
Wu, Jiajun
Fei-Fei, Li
Li, Manling
author_facet Wang, Qineng
Huang, Wenlong
Zhou, Yu
Yin, Hang
Bao, Tianwei
Lyu, Jianwen
Liu, Weiyu
Zhang, Ruohan
Wu, Jiajun
Fei-Fei, Li
Li, Manling
contents Embodied cognition argues that intelligence arises from sensorimotor interaction rather than passive observation. It raises an intriguing question: do modern vision-language models (VLMs), trained largely in a disembodied manner, exhibit signs of embodied cognition? We introduce ENACT, a benchmark that casts evaluation of embodied cognition as world modeling from egocentric interaction in a visual question answering (VQA) format. Framed as a partially observable Markov decision process (POMDP) whose actions are scene graph changes, ENACT comprises two complementary sequence reordering tasks: forward world modeling (reorder shuffled observations given actions) and inverse world modeling (reorder shuffled actions given observations). While conceptually simple, solving these tasks implicitly demands capabilities central to embodied cognition-affordance recognition, action-effect reasoning, embodied awareness, and interactive, long-horizon memory from partially observable egocentric input, while avoiding low-level image synthesis that could confound the evaluation. We provide a scalable pipeline that synthesizes QA pairs from robotics simulation (BEHAVIOR) and evaluates models on 8,972 QA pairs spanning long-horizon home-scale activities. Experiments reveal a performance gap between frontier VLMs and humans that widens with interaction horizon. Models consistently perform better on the inverse task than the forward one and exhibit anthropocentric biases, including a preference for right-handed actions and degradation when camera intrinsics or viewpoints deviate from human vision. Website at https://enact-embodied-cognition.github.io/.
format Preprint
id arxiv_https___arxiv_org_abs_2511_20937
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ENACT: Evaluating Embodied Cognition with World Modeling of Egocentric Interaction
Wang, Qineng
Huang, Wenlong
Zhou, Yu
Yin, Hang
Bao, Tianwei
Lyu, Jianwen
Liu, Weiyu
Zhang, Ruohan
Wu, Jiajun
Fei-Fei, Li
Li, Manling
Artificial Intelligence
Computation and Language
Computer Vision and Pattern Recognition
Robotics
Embodied cognition argues that intelligence arises from sensorimotor interaction rather than passive observation. It raises an intriguing question: do modern vision-language models (VLMs), trained largely in a disembodied manner, exhibit signs of embodied cognition? We introduce ENACT, a benchmark that casts evaluation of embodied cognition as world modeling from egocentric interaction in a visual question answering (VQA) format. Framed as a partially observable Markov decision process (POMDP) whose actions are scene graph changes, ENACT comprises two complementary sequence reordering tasks: forward world modeling (reorder shuffled observations given actions) and inverse world modeling (reorder shuffled actions given observations). While conceptually simple, solving these tasks implicitly demands capabilities central to embodied cognition-affordance recognition, action-effect reasoning, embodied awareness, and interactive, long-horizon memory from partially observable egocentric input, while avoiding low-level image synthesis that could confound the evaluation. We provide a scalable pipeline that synthesizes QA pairs from robotics simulation (BEHAVIOR) and evaluates models on 8,972 QA pairs spanning long-horizon home-scale activities. Experiments reveal a performance gap between frontier VLMs and humans that widens with interaction horizon. Models consistently perform better on the inverse task than the forward one and exhibit anthropocentric biases, including a preference for right-handed actions and degradation when camera intrinsics or viewpoints deviate from human vision. Website at https://enact-embodied-cognition.github.io/.
title ENACT: Evaluating Embodied Cognition with World Modeling of Egocentric Interaction
topic Artificial Intelligence
Computation and Language
Computer Vision and Pattern Recognition
Robotics
url https://arxiv.org/abs/2511.20937