WoMAP: World Models For Embodied Open-Vocabulary Object Localization

TL;DR

WoMAP introduces a scalable, open-vocabulary object localization method for robots that combines world models, real-to-sim data generation, and dense reward distillation, achieving superior zero-shot performance and robust sim-to-real transfer.

Contribution

It presents a novel pipeline integrating Gaussian Splatting, dense reward distillation, and latent world models for open-vocabulary object localization without expert demonstrations.

Findings

Over 9x higher success rate than VLM baseline.

More than 2x success rate compared to diffusion policy.

Effective zero-shot generalization and sim-to-real transfer.

Abstract

Language-instructed active object localization is a critical challenge for robots, requiring efficient exploration of partially observable environments. However, state-of-the-art approaches either struggle to generalize beyond demonstration datasets (e.g., imitation learning methods) or fail to generate physically grounded actions (e.g., VLMs). To address these limitations, we introduce WoMAP (World Models for Active Perception): a recipe for training open-vocabulary object localization policies that: (i) uses a Gaussian Splatting-based real-to-sim-to-real pipeline for scalable data generation without the need for expert demonstrations, (ii) distills dense rewards signals from open-vocabulary object detectors, and (iii) leverages a latent world model for dynamics and rewards prediction to ground high-level action proposals at inference time. Rigorous simulation and hardware experiments demonstrate WoMAP's superior performance in a broad range of zero-shot object localization tasks, with more than 9x and 2x higher success rates compared to VLM and diffusion policy baselines, respectively. Further, we show that WoMAP achieves strong generalization and sim-to-real transfer on a TidyBot.