# T2-z：机器人世界模型的环境潜变量测试时优化

> 原文链接：https://robotics.hellobike.com/papers/2026-10-15-2610-22364
> 频道：papers | 更新日期：2026-10-15 | 标签：arxiv, cs.RO
> 核心摘要：T2-Z通过测试时优化可学习环境潜变量，使机器人世界模型适应未见环境，并在仿真设置和真实机器人任务上验证。

# T2-z：机器人世界模型的环境潜变量测试时优化

::: v-pre
> 原文链接：[T2-z: Test-Time Optimization of Environment Latents for Robotic World Models](https://arxiv.org/abs/2610.22364)（arXiv 2610.22364 · 发布日期：2026-10-15）

**T2-z: Test-Time Optimization of Environment Latents for Robotic World Models**

T2-Z通过测试时优化可学习环境潜变量，使机器人世界模型适应未见环境，并在仿真设置和真实机器人任务上验证。

## 核心要点

- 机器人预测动作后果需考虑不可见、未显式指定的摩擦、质量或环境规则。
- T2-Z为世界模型加入可学习环境潜变量，交替优化动力学模型与条件潜变量。
- 测试时冻结世界模型，用一次或多次交互优化新潜变量以适应未见环境配置。
- 7个仿真和4个真实任务中，误差15.20降至7.00，成功率33.73%升至72.96%。

**验证范围与局限**：验证范围为7个模拟设置和4个真实世界机器人任务。

## 英文摘要 (Abstract)

Predicting how the world responds to an action requires more than what can be seen: robots must account for environmental factors that are neither directly observable nor explicitly specified, from physical properties such as friction and mass to human-defined environmental rules. We introduce T2-Z for test-time optimization of environment latents, which augments a robotic world model with a learnable environment latent to condition world predictions. This compact latent captures underlying environmental variation without supervision of what environmental factors exist or what values they take. During training, T2-Z alternates between two stages: a world-model optimization stage that fixes the latents and learns general action-conditioned dynamics across environments; and an environment-latent optimization stage that freezes the world model and optimizes the conditioning latents, allowing latent structure to emerge purely through world prediction. At test time, the world model remains frozen and a newly initialized environment latent is optimized from one or more test-time interactions, adapting the model to previously unseen environmental configurations. Across 7 simulated settings and 4 real-world robotic tasks, T2-Z reduces object prediction error metric from 15.20 to 7.00 and raises action success from 33.73% to 72.96%. Its emergent latents capture structured representations of individual and compounded factors across diverse visual scenes. Ablations further show that T2-Z can efficiently infer environment latents from as little as a single test-time interaction trajectory.

作者：Chenyuan Zhou, Yining Hong, Huang Huang, Yunfan Jiang, Wenlong Huang, David Held, Vitor Guizilini, Yue Wang, Li Fei-Fei, Jiajun Wu
:::

## 原始参考来源 (Sources)

- [arXiv 2610.22364：T2-z: Test-Time Optimization of Environment Latents for Robotic World Models](https://arxiv.org/abs/2610.22364)