Probability Functional Descent: A Unifying Perspective on GANs, Variational Inference, and Reinforcement Learning

Probability Functional Descent: A Unifying Perspective on GANs, Variational Inference, and Reinforcement Learning
复制标题

DOI：
--
复制
发表时间：
2019-01
期刊：
ArXiv
影响因子：
--
通讯作者：
Casey Chu;J. Blanchet;P. Glynn
Casey Chu;J. Blanchet;P. Glynn
中科院分区：
其他
文献类型：
--
作者：
Casey Chu;J. Blanchet;P. Glynn

文献摘要

被引文献

相似文献

本文通过将机器学习中广泛关注的问题定义为概率测度空间上定义的泛函的最小化，提供了对这些问题的统一观点。特别是，我们表明，强化学习中的生成对抗网络、变分推理和演员批评家方法都可以通过我们的框架的镜头看到。然后，我们讨论我们的公式的通用优化算法，称为概率函数下降（PFD），并展示该算法如何恢复在前面提到的设置中独立开发的现有方法。
This paper provides a unifying view of a wide range of problems of interest in machine learning by framing them as the minimization of functionals defined on the space of probability measures. In particular, we show that generative adversarial networks, variational inference, and actor-critic methods in reinforcement learning can all be seen through the lens of our framework. We then discuss a generic optimization algorithm for our formulation, called probability functional descent (PFD), and show how this algorithm recovers existing methods developed independently in the settings mentioned earlier.