arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2502.01366cs.LG
收藏

面向异构环境的轨迹世界模型

Trajectory World Models for Heterogeneous Environments

Tsinghua University(清华大学)

Shaofeng Yin, Jialong Wu, Siqiao Huang, Xingjian Su, Xu He, Jianye Hao, Mingsheng Long

AI总结:

针对传感器与执行器异构导致的世界模型迁移难题,本文构建含80个环境、超百万轨迹的UniTraj数据集,并提出可灵活适配异构信息、上下文建模动态的TrajWorld架构,显著提升预测、离策略评估与MPC在线性能。

摘要:

跨环境的传感器与执行器异构性,给基于这类低维传感器信息构建大规模预训练世界模型带来了重大挑战。在本工作中,我们通过解决数据多样性和模型灵活性两方面的关键迁移障碍,探索为异构环境预训练世界模型。我们介绍了 UniTraj,这是一个统一数据集,包含来自 80 个环境的超过一百万条轨迹,旨在扩展数据规模的同时保留关键多样性。此外,我们提出了 TrajWorld,这是一种新颖架构,能够灵活处理变化的传感器和执行器信息,并在上下文中捕捉环境动态。在 UniTraj 上预训练 TrajWorld 可在转移动态预测方面取得显著提升,在离策略评估上达到新的最先进水平,并且在模型预测控制的在线性能上也表现更优。据我们所知,本工作首次证明了世界模型在异构且复杂的控制环境之间的迁移收益。代码和数据可在 https://github.com/thuml/TrajWorld 获取。

英文摘要:

Heterogeneity in sensors and actuators across environments poses a significant challenge to building large-scale pre-trained world models on top of this low-dimensional sensor information. In this work, we explore pre-training world models for heterogeneous environments by addressing key transfer barriers in both data diversity and model flexibility. We introduce UniTraj, a unified dataset comprising over one million trajectories from 80 environments, designed to scale data while preserving critical diversity. Additionally, we propose TrajWorld, a novel architecture capable of flexibly handling varying sensor and actuator information and capturing environment dynamics in-context. Pre-training TrajWorld on UniTraj yields substantial gains in transition prediction, achieves a new state-of-the-art for off-policy evaluation, and also delivers superior online performance of model predictive control. To the best of our knowledge, this work, for the first time, demonstrates the transfer benefits of world models across heterogeneous and complex control environments. Code and data are available at https://github.com/thuml/TrajWorld.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑