VisualPredicator: Learning Abstract World Models with Neuro-Symbolic Predicates for Robot Planning

TL;DR

This paper introduces Neuro-Symbolic Predicates, a novel approach combining neural and symbolic representations to create task-specific abstractions for robot planning, improving generalization, interpretability, and sample efficiency.

Contribution

It proposes a new online algorithm for inventing and learning abstract world models using Neuro-Symbolic Predicates, advancing robot planning capabilities.

Findings

Better sample complexity compared to existing methods

Stronger out-of-distribution generalization

Enhanced interpretability of learned models

Abstract

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.