Sparse Imagination for Efficient Visual World Model Planning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chun, Junha, Jeong, Youngjoon, Kim, Taesup
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914351229698048
author Chun, Junha
Jeong, Youngjoon
Kim, Taesup
author_facet Chun, Junha
Jeong, Youngjoon
Kim, Taesup
contents World model based planning has significantly improved decision-making in complex environments by enabling agents to simulate future states and make informed choices. This computational burden is particularly restrictive in robotics, where resources are severely constrained. To address this limitation, we propose a Sparse Imagination for Efficient Visual World Model Planning, which enhances computational efficiency by reducing the number of tokens processed during forward prediction. Our method leverages a sparsely trained vision-based world model based on transformers with randomized grouped attention strategy, allowing the model to flexibly adjust the number of tokens processed based on the computational resource. By enabling sparse imagination during latent rollout, our approach significantly accelerates planning while maintaining high control fidelity. Experimental results demonstrate that sparse imagination preserves task performance while dramatically improving inference efficiency. This general technique for visual planning is applicable from simple test-time trajectory optimization to complex real-world tasks with the latest VLAs, enabling the deployment of world models in real-time scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2506_01392
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sparse Imagination for Efficient Visual World Model Planning
Chun, Junha
Jeong, Youngjoon
Kim, Taesup
Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
World model based planning has significantly improved decision-making in complex environments by enabling agents to simulate future states and make informed choices. This computational burden is particularly restrictive in robotics, where resources are severely constrained. To address this limitation, we propose a Sparse Imagination for Efficient Visual World Model Planning, which enhances computational efficiency by reducing the number of tokens processed during forward prediction. Our method leverages a sparsely trained vision-based world model based on transformers with randomized grouped attention strategy, allowing the model to flexibly adjust the number of tokens processed based on the computational resource. By enabling sparse imagination during latent rollout, our approach significantly accelerates planning while maintaining high control fidelity. Experimental results demonstrate that sparse imagination preserves task performance while dramatically improving inference efficiency. This general technique for visual planning is applicable from simple test-time trajectory optimization to complex real-world tasks with the latest VLAs, enabling the deployment of world models in real-time scenarios.
title Sparse Imagination for Efficient Visual World Model Planning
topic Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.01392