VisualPredicator: Learning Abstract World Models with Neuro-Symbolic Predicates for Robot Planning
This work addresses the challenge of abstraction for robot planning, offering incremental improvements in generalization and efficiency for simulated robotic domains.
The paper tackled the problem of enabling intelligent agents to form task-specific abstractions by introducing Neuro-Symbolic Predicates, a first-order abstraction language combining symbolic and neural representations, with results showing better sample complexity, stronger out-of-distribution generalization, and improved interpretability compared to existing methods.
Broadly intelligent agents should form task-specific abstractions that selectively expose the essential elements of a task, while abstracting away the complexity of the raw sensorimotor space. In this work, we present Neuro-Symbolic Predicates, a first-order abstraction language that combines the strengths of symbolic and neural knowledge representations. We outline an online algorithm for inventing such predicates and learning abstract world models. We compare our approach to hierarchical reinforcement learning, vision-language model planning, and symbolic predicate invention approaches, on both in- and out-of-distribution tasks across five simulated robotic domains. Results show that our approach offers better sample complexity, stronger out-of-distribution generalization, and improved interpretability.