Interpretable concept world model CGWM: a new neurosymbolic planning approach
What happened
Researcher Riccardo Andreoni has proposed the Concept-Grounded World Model (CGWM) on arXiv. It is a neurosymbolic design that learns interpretable concept representations from observations. The model learns from knowledge of valid world states and their dynamics, expressed in temporal logic, and it frames planning as differentiable reasoning, optimizing the agent's actions by gradient ascent to meet a goal. In experiments in two game environments, the paper compares CGWM with a purely neural baseline and a neurosymbolic baseline. The authors report that CGWM's concept grounding is markedly more accurate, including over long horizons, and that its planning is more robust both in and out of distribution. These results are self-reported in a preprint that has not yet been peer-reviewed or independently verified.
AI-written from the coverage · updated 2026-10-15
Coverage timeline
Follow the coverage to see the story from different sides.
- arXiv Game World Models & GenerationNeurosymbolic learning of interpretable world models
Researchers propose the Concept-Grounded World Model (CGWM), a neurosymbolic design that learns interpretable concept representations from observations. It is built on knowledge of valid world states and their dynamics, expressed in temporal logic, and it frames planning as differentiable reasoning, optimizing the agent's actions by gradient ascent to meet a goal. In experiments in two game environments, CGWM's concept grounding is markedly more accurate, including over long horizons, and its planning is more robust both in and out of distribution.