T2-Z：测试时优化环境隐变量，提升机器人世界模型预测能力
T2-z: Test-Time Optimization of Environment Latents for Robotic World Models
这篇 T2-Z 挺有意思，给机器人世界模型加了个环境隐变量，测试时优化一下就能适应新环境，动作成功率从 34% 拉到 73%，做机器人方向的朋友可以看看。
T2-Z 是一种面向机器人世界模型的测试时优化方法，通过一个可学习的环境隐变量来条件化世界预测，无需监督即可捕捉摩擦、质量等不可直接观测的环境因素。训练时交替进行两个阶段：固定隐变量优化世界模型，再冻结世界模型优化环境隐变量。测试时世界模型保持冻结，仅用一次或多次交互轨迹优化新初始化的环境隐变量，即可适配未见过的环境配置。在 7 个模拟设置和 4 个真实机器人任务上，T2-Z 将物体预测误差从 15.20 降至 7.00，动作成功率从 33.73% 提升至 72.96%。
T2-z: Test-Time Optimization of Environment Latents for Robotic World Models
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.