Learning to Theorize the World from Observation

2026-05-05

AI Summary

This work addresses the challenge of constructing interpretable world understanding from raw, non-textual observations rather than merely predicting future states. It introduces a novel “learning-to-theorize” paradigm that, for the first time, incorporates theory-building mechanisms from developmental cognitive science into machine learning. The approach employs a Neural Episodic Theorizer (NEO) to induce explicit, executable, program-like world theories from data. By integrating probabilistic neural models, implicit program induction, shared transition dynamics, and a learned Language of Thought, the framework represents theories as composable programs, thereby enabling systematic generalization. Experiments demonstrate that the model achieves explanation-driven generalization by leveraging the underlying generative programs of observed phenomena, effectively supporting the comprehension of novel situations.

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.

Problem

• world understanding
• theory learning
• explanatory theories
• non-textual observations
• cognitive modeling

Innovation

• Learning-to-Theorize
• Neural Theorizer
• Language of Thought
• executable programs
• explanation-driven generalization