Invariance Through Latent Alignment
Invariance Through Latent Alignment
复制标题
DOI：
10.15607/rss.2022.xviii.064
复制
发表时间：
2021-12
期刊：
影响因子：
--
通讯作者：
Takuma Yoneda;Ge Yang;Matthew R. Walter;Bradly C. Stadie
中科院分区：
文献类型：
--
作者：
Takuma Yoneda;Ge Yang;Matthew R. Walter;Bradly C. Stadie
机器人的部署环境通常会涉及与训练期间所经历的不同的感知变化。数据增强等标准实践试图通过增强源图像来弥补这一差距，以扩展对训练分布的支持，以更好地覆盖代理在测试时可能遇到的情况。然而，在许多情况下，不可能先验地知道测试时间分布转移，使得这些方案不可行。在本文中，我们介绍了一种称为“潜在对齐不变性”（ILA）的通用方法，该方法可以提高视觉运动控制策略在感知变化未知的部署环境中的测试时性能。 ILA 在部署时通过将目标域上的潜在特征的分布与代理的先前经验相匹配来执行无监督的适应，而不依赖于配对数据。虽然很简单，但我们证明这个想法可以在各种具有挑战性的适应场景中带来令人惊讶的改进，包括照明条件、场景中的内容和相机姿势的变化。我们展示了模拟中校准控制基准的结果——干扰器控制套件——以及模拟到真实设置下的物理机器人。
A robot's deployment environment often involves perceptual changes that differ from what it has experienced during training. Standard practices such as data augmentation attempt to bridge this gap by augmenting source images in an effort to extend the support of the training distribution to better cover what the agent might experience at test time. In many cases, however, it is impossible to know test-time distribution-shift a priori, making these schemes infeasible. In this paper, we introduce a general approach, called Invariance Through Latent Alignment (ILA), that improves the test-time performance of a visuomotor control policy in deployment environments with unknown perceptual variations. ILA performs unsupervised adaptation at deployment-time by matching the distribution of latent features on the target domain to the agent's prior experience, without relying on paired data. Although simple, we show that this idea leads to surprising improvements on a variety of challenging adaptation scenarios, including changes in lighting conditions, the content in the scene, and camera poses. We present results on calibrated control benchmarks in simulation -- the distractor control suite -- and a physical robot under a sim-to-real setup.