Learning to Theorize the World from Observation

Doojin Baek, Gyubin Lee, Junyeob Baek, Hosung Lee, Sungjin Ahn
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:5064-5106, 2026.

Abstract

What does it mean to understand the world? Contemporary world models often operationalize understanding as accurate future prediction in latent or observation space. Developmental cognitive science, however, suggests a different view: human understanding emerges through the construction of internal theories of how the world works, even before mature language is acquired. Inspired by this theory-building view of cognition, we introduce Learning-to-Theorize, a learning paradigm for inferring explicit explanatory theories of the world from raw, non-textual observations. We instantiate this paradigm with the Neural Theorizer (NEO), a World Theory Model, that induces latent programs as a learned Language of Thought and executes them through a shared transition model. In NEO, a theory is represented as an executable, compositional program whose learned primitives can be systematically recombined to explain novel phenomena. Experiments show that this formulation enables explanation-driven generalization, allowing observations to be understood in terms of the programs that generate them. Our code is available at github.com/ahn-ml/learning-to-theorize

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-baek26b, title = {Learning to Theorize the World from Observation}, author = {Baek, Doojin and Lee, Gyubin and Baek, Junyeob and Lee, Hosung and Ahn, Sungjin}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {5064--5106}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/baek26b/baek26b.pdf}, url = {https://proceedings.mlr.press/v306/baek26b.html}, abstract = {What does it mean to understand the world? Contemporary world models often operationalize understanding as accurate future prediction in latent or observation space. Developmental cognitive science, however, suggests a different view: human understanding emerges through the construction of internal theories of how the world works, even before mature language is acquired. Inspired by this theory-building view of cognition, we introduce Learning-to-Theorize, a learning paradigm for inferring explicit explanatory theories of the world from raw, non-textual observations. We instantiate this paradigm with the Neural Theorizer (NEO), a World Theory Model, that induces latent programs as a learned Language of Thought and executes them through a shared transition model. In NEO, a theory is represented as an executable, compositional program whose learned primitives can be systematically recombined to explain novel phenomena. Experiments show that this formulation enables explanation-driven generalization, allowing observations to be understood in terms of the programs that generate them. Our code is available at github.com/ahn-ml/learning-to-theorize} }
Endnote
%0 Conference Paper %T Learning to Theorize the World from Observation %A Doojin Baek %A Gyubin Lee %A Junyeob Baek %A Hosung Lee %A Sungjin Ahn %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-baek26b %I PMLR %P 5064--5106 %U https://proceedings.mlr.press/v306/baek26b.html %V 306 %X What does it mean to understand the world? Contemporary world models often operationalize understanding as accurate future prediction in latent or observation space. Developmental cognitive science, however, suggests a different view: human understanding emerges through the construction of internal theories of how the world works, even before mature language is acquired. Inspired by this theory-building view of cognition, we introduce Learning-to-Theorize, a learning paradigm for inferring explicit explanatory theories of the world from raw, non-textual observations. We instantiate this paradigm with the Neural Theorizer (NEO), a World Theory Model, that induces latent programs as a learned Language of Thought and executes them through a shared transition model. In NEO, a theory is represented as an executable, compositional program whose learned primitives can be systematically recombined to explain novel phenomena. Experiments show that this formulation enables explanation-driven generalization, allowing observations to be understood in terms of the programs that generate them. Our code is available at github.com/ahn-ml/learning-to-theorize
APA
Baek, D., Lee, G., Baek, J., Lee, H. & Ahn, S.. (2026). Learning to Theorize the World from Observation. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:5064-5106 Available from https://proceedings.mlr.press/v306/baek26b.html.

Related Material