MoWM: Mixture-of-World-Models for Embodied Planning via Latent-to-Pixel Feature Modulation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yu, Yangcheng, Jin, Xin, Shang, Yu, Zhang, Xin, Su, Haisheng, Wu, Wei, Li, Yong
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911435992334336
author Yu, Yangcheng
Jin, Xin
Shang, Yu
Zhang, Xin
Su, Haisheng
Wu, Wei
Li, Yong
author_facet Yu, Yangcheng
Jin, Xin
Shang, Yu
Zhang, Xin
Su, Haisheng
Wu, Wei
Li, Yong
contents Embodied action planning is a core challenge in robotics, requiring models to generate precise actions from visual observations and language instructions. While video generation world models are promising, their reliance on pixel-level reconstruction often introduces visual redundancies that hinder action decoding and generalization. Latent world models offer a compact, motion-aware representation, but overlook the fine-grained details critical for precise manipulation. To overcome these limitations, we propose MoWM, a mixture-of-world-model framework that fuses representations from hybrid world models for embodied action planning. Our approach combines motion-aware latent world model features with pixel-space features, enabling MoWM to emphasize action-relevant visual details for action decoding. Extensive evaluations on the CALVIN and real-world manipulation tasks demonstrate that our method achieves state-of-the-art task success rates and superior generalization. We also provide a comprehensive analysis of the strengths of each feature space, offering valuable insights for future research in embodied planning. The code is available at: https://github.com/tsinghua-fib-lab/MoWM.
format Preprint
id arxiv_https___arxiv_org_abs_2509_21797
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MoWM: Mixture-of-World-Models for Embodied Planning via Latent-to-Pixel Feature Modulation
Yu, Yangcheng
Jin, Xin
Shang, Yu
Zhang, Xin
Su, Haisheng
Wu, Wei
Li, Yong
Computer Vision and Pattern Recognition
Embodied action planning is a core challenge in robotics, requiring models to generate precise actions from visual observations and language instructions. While video generation world models are promising, their reliance on pixel-level reconstruction often introduces visual redundancies that hinder action decoding and generalization. Latent world models offer a compact, motion-aware representation, but overlook the fine-grained details critical for precise manipulation. To overcome these limitations, we propose MoWM, a mixture-of-world-model framework that fuses representations from hybrid world models for embodied action planning. Our approach combines motion-aware latent world model features with pixel-space features, enabling MoWM to emphasize action-relevant visual details for action decoding. Extensive evaluations on the CALVIN and real-world manipulation tasks demonstrate that our method achieves state-of-the-art task success rates and superior generalization. We also provide a comprehensive analysis of the strengths of each feature space, offering valuable insights for future research in embodied planning. The code is available at: https://github.com/tsinghua-fib-lab/MoWM.
title MoWM: Mixture-of-World-Models for Embodied Planning via Latent-to-Pixel Feature Modulation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.21797