ExoPredicator: Learning Abstract Models of Dynamic Worlds for Robot Planning
Abstract
Long-horizon embodied planning is challenging because the world does not only change through an agent's actions: exogenous processes (e.g., water heating, dominoes cascading) unfold concurrently with the agent's actions. We propose a framework for abstract world models that jointly learns (i) symbolic state representations and (ii) causal processes for both endogenous actions and exogenous mechanisms. Each causal process models the time course of a stochastic cause-effect relation. We learn these world models from limited data via variational Bayesian inference combined with LLM proposals. Across five simulated tabletop robotics environments, the learned models enable fast planning that generalizes to held-out tasks with more objects and more complex goals, outperforming a range of baselines.
Cite
Text
Liang et al. "ExoPredicator: Learning Abstract Models of Dynamic Worlds for Robot Planning." International Conference on Learning Representations, 2026.Markdown
[Liang et al. "ExoPredicator: Learning Abstract Models of Dynamic Worlds for Robot Planning." International Conference on Learning Representations, 2026.](https://mlanthology.org/iclr/2026/liang2026iclr-exopredicator/)BibTeX
@inproceedings{liang2026iclr-exopredicator,
title = {{ExoPredicator: Learning Abstract Models of Dynamic Worlds for Robot Planning}},
author = {Liang, Yichao and Nguyen, Thanh Dat and Yang, Cambridge and Li, Tianyang and Tenenbaum, Joshua B. and Rasmussen, Carl Edward and Weller, Adrian and Tavares, Zenna and Silver, Tom and Ellis, Kevin},
booktitle = {International Conference on Learning Representations},
year = {2026},
url = {https://mlanthology.org/iclr/2026/liang2026iclr-exopredicator/}
}