Saved in:
Bibliographic Details
Main Authors: Parés-Morlans, Carlota, Yi, Michelle, Chen, Claire, Wu, Sarah A., Antonova, Rika, Gerstenberg, Tobias, Bohg, Jeannette
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.22861
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916766871977984
author Parés-Morlans, Carlota
Yi, Michelle
Chen, Claire
Wu, Sarah A.
Antonova, Rika
Gerstenberg, Tobias
Bohg, Jeannette
author_facet Parés-Morlans, Carlota
Yi, Michelle
Chen, Claire
Wu, Sarah A.
Antonova, Rika
Gerstenberg, Tobias
Bohg, Jeannette
contents Tasks that involve complex interactions between objects with unknown dynamics make planning before execution difficult. These tasks require agents to iteratively improve their actions after actively exploring causes and effects in the environment. For these type of tasks, we propose Causal-PIK, a method that leverages Bayesian optimization to reason about causal interactions via a Physics-Informed Kernel to help guide efficient search for the best next action. Experimental results on Virtual Tools and PHYRE physical reasoning benchmarks show that Causal-PIK outperforms state-of-the-art results, requiring fewer actions to reach the goal. We also compare Causal-PIK to human studies, including results from a new user study we conducted on the PHYRE benchmark. We find that Causal-PIK remains competitive on tasks that are very challenging, even for human problem-solvers.
format Preprint
id arxiv_https___arxiv_org_abs_2505_22861
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Causal-PIK: Causality-based Physical Reasoning with a Physics-Informed Kernel
Parés-Morlans, Carlota
Yi, Michelle
Chen, Claire
Wu, Sarah A.
Antonova, Rika
Gerstenberg, Tobias
Bohg, Jeannette
Machine Learning
Tasks that involve complex interactions between objects with unknown dynamics make planning before execution difficult. These tasks require agents to iteratively improve their actions after actively exploring causes and effects in the environment. For these type of tasks, we propose Causal-PIK, a method that leverages Bayesian optimization to reason about causal interactions via a Physics-Informed Kernel to help guide efficient search for the best next action. Experimental results on Virtual Tools and PHYRE physical reasoning benchmarks show that Causal-PIK outperforms state-of-the-art results, requiring fewer actions to reach the goal. We also compare Causal-PIK to human studies, including results from a new user study we conducted on the PHYRE benchmark. We find that Causal-PIK remains competitive on tasks that are very challenging, even for human problem-solvers.
title Causal-PIK: Causality-based Physical Reasoning with a Physics-Informed Kernel
topic Machine Learning
url https://arxiv.org/abs/2505.22861