构建跨异构环境的统一世界模型,提升复杂控制任务的泛化能力。
Trajectory World Models for Heterogeneous Environments
- 提出UniTraj数据集与TrajWorld架构,支持多传感器多动作输入。
- 在百万级轨迹上预训练后,预测准确率与离线评估性能达新高。
- 适用于机器人控制、强化学习等需要跨环境迁移的场景。
不同环境中的传感器和执行器存在异质性,给基于低维传感信息构建大规模预训练世界模型带来挑战。本文通过解决数据多样性和模型灵活性两大转移障碍,探索异构环境下的世界模型预训练。我们提出了UniTraj,一个包含超过一百万条轨迹的统一数据集,覆盖80个不同环境,兼顾规模与关键多样性。同时提出TrajWorld,一种新型架构,能灵活处理变化的传感器与执行器信息,并在上下文中捕捉环境动态。在UniTraj上预训练TrajWorld后,其在状态转移预测上取得显著提升,实现离线评估的新最佳表现,并在模型预测控制的在线性能上也优于现有方法。据我们所知,这是首次在异构复杂控制环境中验证世界模型的迁移优势。代码与数据可在https://github.com/thuml/TrajWorld获取。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。