Temporal Difference Learning: Why Updating a Guess With a Guess Actually Works
中文摘要
本文深入浅出地解释了现代强化学习的核心——时间差分学习（TD Learning），阐述了通过预测更新预测的运作原理。
English Summary
This article explains Temporal Difference Learning, a fundamental concept of modern reinforcement learning that updates predictions based on other predictions, without using any code.
原文节选
The one idea holding up most of modern reinforcement learning — explained without a single line of code Continue reading on Medium »