Extracting Visual Plans from Unlabeled Videos via Symbolic Guidance

TL;DR

Vis2Plan is a symbolic, explainable visual planning framework that extracts task symbols from unlabeled videos, enabling efficient, high-success-rate multi-goal planning for robotic manipulation.

Contribution

It introduces a white-box, symbolic approach to visual planning that outperforms diffusion-based methods in success rate and speed, using vision foundation models for symbol extraction.

Findings

53% higher success rate than diffusion-based planners

35 times faster visual plan generation

Physically consistent and inspectable visual plans

Abstract

Visual planning, by offering a sequence of intermediate visual subgoals to a goal-conditioned low-level policy, achieves promising performance on long-horizon manipulation tasks. To obtain the subgoals, existing methods typically resort to video generation models but suffer from model hallucination and computational cost. We present Vis2Plan, an efficient, explainable and white-box visual planning framework powered by symbolic guidance. From raw, unlabeled play data, Vis2Plan harnesses vision foundation models to automatically extract a compact set of task symbols, which allows building a high-level symbolic transition graph for multi-goal, multi-stage planning. At test time, given a desired task goal, our planner conducts planning at the symbolic level and assembles a sequence of physically consistent intermediate sub-goal images grounded by the underlying symbolic representation. Our Vis2Plan outperforms strong diffusion video generation-based visual planners by delivering 53\% higher aggregate success rate in real robot settings while generating visual plans 35$\times$ faster. The results indicate that Vis2Plan is able to generate physically consistent image goals while offering fully inspectable reasoning steps.