BriefGPT.xyz
Feb, 2019
全随机梯度算法及其在强化学习中的应用
Total stochastic gradient algorithms and applications in reinforcement learning
HTML
PDF
Paavo Parmas
TL;DR
本文介绍了如何利用总导数规则创建图模型的梯度估算器，并基于密度估计和似然比梯度推导了新的梯度估算器。通过在基于模型的策略梯度算法中测试，本文证明了这些方法的有效性，并揭示了PILCO算法的成功之谜。
Abstract
backpropagation
and the chain rule of derivatives have been prominent; however, the
total derivative rule
has not enjoyed the same amount of attention. In this work we show how the
→