STRIPS-WM: Learning Grounded Propositional STRIPS-style World Models from Images

2026-06-04

AI Summary

This work addresses the challenge of automatically learning symbolic action models from high-dimensional images for classical task planning in long-horizon visual manipulation. Leveraging image-based state transition data, the method constructs an abstract state graph and jointly learns binary predicates aligned with visual observations alongside propositional STRIPS operators for corresponding actions. Through knowledge distillation, predicate semantics are injected into the visual encoder, enabling an end-to-end mapping from raw images to symbolic plans. This approach is the first to automatically induce STRIPS-style action models—with sparse preconditions and add/delete effects—solely from visual experience. Evaluated on visual rearrangement tasks, it significantly outperforms baselines based on visual rollout, latent-space graph search, and latent symbolic representations, achieving markedly higher success rates in translating images into executable plans.

Abstract

Robots performing long-horizon visual manipulation observe high-dimensional images, but successful plans depend on action-relevant facts: what can be done now and what changes afterward. A useful planning representation should discard irrelevant visual details while preserving action applicability and effects. Classical task planners exploit this structure through symbolic operators with preconditions and effects, but obtaining such representations from raw visual experience remains challenging. We study a visual task-planning setting in which a robot receives only image transitions: the current image, executed high-level action, and the resulting image. At test time, given a start image and a goal image, the robot must produce a sequence of high-level actions that reaches the goal. To address this problem, we introduce STRIPS-WM, a framework for learning image-grounded STRIPS-style world models directly from visual transitions. STRIPS-WM first induces a finite abstract transition graph from images, then learns latent binary predicates and one grounded propositional operator per action label. The learned operators form a symbolic action model with sparse preconditions and add/delete effects. Finally, the learned predicates are distilled into a visual encoder, enabling classical planning directly from novel start and goal images. Experiments on visual rearrangement tasks show that STRIPS-WM improves image-to-plan success over the tested visual rollout, latent graph-search and latent-symbolic baselines.

Problem

• visual planning
• symbolic world models
• STRIPS
• image-to-plan
• action representation

Innovation

• STRIPS-style planning
• world model learning
• visual abstraction
• symbolic action models
• predicate distillation