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 2111.05458stat.MLcs.LG
收藏

哪些前科很重要?学习潜在动态的基准模型

Which priors matter? Benchmarking models for learning latent dynamics

DeepMind

Aleksandar Botev, Andrew Jaegle, Peter Wirnsberger, Daniel Hennes, Irina Higgins

摘要:

学习动态是机器学习(ML)许多重要应用的核心,例如机器人和自动驾驶。在这些设置中,ML算法通常需要使用高维观测(如图像)来推理物理系统,而无需访问底层状态。最近,有几种方法提出将经典力学的先验知识集成到ML模型中,以解决从图像进行物理推理的挑战。在这项工作中,我们冷静地看待这些模型的当前能力。为此,我们介绍了一套由17个数据集组成的可视化观察的基础上表现出广泛的动态物理系统。我们对主要的物理启发方法进行了全面而详细的比较,并与几个强大的基线进行了比较。虽然包含物理先验的模型通常可以学习具有所需属性的潜在空间,但我们的结果表明,这些方法未能显着改善标准技术。尽管如此,我们发现,使用连续和时间可逆的动力学有利于所有类别的模型。

英文摘要:

Learning dynamics is at the heart of many important applications of machine learning (ML), such as robotics and autonomous driving. In these settings, ML algorithms typically need to reason about a physical system using high dimensional observations, such as images, without access to the underlying state. Recently, several methods have proposed to integrate priors from classical mechanics into ML models to address the challenge of physical reasoning from images. In this work, we take a sober look at the current capabilities of these models. To this end, we introduce a suite consisting of 17 datasets with visual observations based on physical systems exhibiting a wide range of dynamics. We conduct a thorough and detailed comparison of the major classes of physically inspired methods alongside several strong baselines. While models that incorporate physical priors can often learn latent spaces with desirable properties, our results demonstrate that these methods fail to significantly improve upon standard techniques. Nonetheless, we find that the use of continuous and time-reversible dynamics benefits models of all classes.

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。

↑