aiqbench analysis
This research proposes a paradigm shift from prediction-based world models to explicit theory construction, utilizing a probabilistic neural model that generates executable programs as a learned Language of Thought. By representing theories as compositional primitives, the approach enables explanation-driven generalization from raw observations, potentially bridging the gap between developmental cognitive science and modern AI architectures. The work suggests a new direction for world models where understanding is defined by the ability to derive and recombine internal causal mechanisms rather than merely forecasting latent states.
Learning to Theorize the World from Observation
arXiv:2605.03413v2 Announce Type: replace-cross 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 probabilistic neural 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.
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 probabilistic neural 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.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2605.03413 [cs.LG]
(or arXiv:2605.03413v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2605.03413 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 245 words
Related coverage
Learning to Theorize the World from Observation
arXiv · Sep 18, 2026
Quantum in the Palm of Your Hand : The Evolution of Superconducting Qubits
newswise.com · Jul 28, 2026
AI excels at exams, but lacks human theory-building skills
Quantum Zeitgeist · Sep 17, 2026
Sympathy for both sides of the egregious misalignment debate
AI Alignment Forum · Jun 12, 2026