arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1807.02128cs.LGcs.ROstat.ML
收藏

自适应路径积分自编码器:动态系统的表示学习与规划

Adaptive Path-Integral Autoencoder: Representation Learning and Planning for Dynamical Systems

KAIST(韩国科学技术院)

Jung-Su Ha, Young-Jin Park, Hyeok-Joo Chae, Soon-Seo Park, Han-Lim Choi

摘要:

我们提出了一种表示学习算法,该算法从高维\textit{sequential}原始数据中学习低维潜在动力系统,例如,视频.该框架建立在摊销推理方法的最新进展,使用推理网络和细化过程,从给定的观察序列的变分分布输出样本,并利用控制和推理之间的对偶性,近似解决棘手的推理问题,使用路径积分控制方法。学习的动态模型可以用来预测和规划未来的状态,我们还提出了有效的规划方法,利用学习的低维潜在的动态。数值实验表明,所提出的基于路径积分控制的变分推理方法,导致更严格的下界在序列数据的统计模型学习。补充视频:https://youtu.be/xCp35crUoLQ

英文摘要:

We present a representation learning algorithm that learns a low-dimensional latent dynamical system from high-dimensional \textit{sequential} raw data, e.g., video. The framework builds upon recent advances in amortized inference methods that use both an inference network and a refinement procedure to output samples from a variational distribution given an observation sequence, and takes advantage of the duality between control and inference to approximately solve the intractable inference problem using the path integral control approach. The learned dynamical model can be used to predict and plan the future states; we also present the efficient planning method that exploits the learned low-dimensional latent dynamics. Numerical experiments show that the proposed path-integral control based variational inference method leads to tighter lower bounds in statistical model learning of sequential data. The supplementary video: https://youtu.be/xCp35crUoLQ

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。

↑