Learn AI Series (#110) - Model-Based Reinforcement Learning What will I learn You will learn the real difference between model-free and model-based RL -- and why the second family exists at all; how to

Learn AI Series (#109) - Advanced Policy Optimization What will I learn You will learn what a trust region is, and why bounding how far a policy may move in one update is the single idea that turns flaky

团委老师在群里突然又发通知，让大家填一个电子表格，还有扫描《入团志愿书》到一个PDF文件，并把表格和PDF文件发给班级团支书。 (插图由ChatGPT辅助生成) 填表格这事是轻车熟路啦，但如何把《入团志愿书》扫描成一个PDF文件，却让我犯了愁，群里很多同学是去复印社花钱搞定，我可舍不得这笔巨款呢。 而且家里的打印机就自带扫描功能，我就不信解决不了这个问题。

Learn AI Series (#108) - Policy Gradient Methods What will I learn You will learn why directly optimizing the policy is sometimes a far better idea than learning a value function first; the policy gradient

Learn AI Series (#107) - Deep Q-Networks (DQN) What will I learn You will learn why tabular Q-Learning falls apart the moment the state space gets big, and why a lookup table was never going to scale to

Learn AI Series (#106) - Temporal Difference Learning What will I learn You will learn TD(0): updating value estimates from single transitions, without ever waiting for an episode to finish; bootstrapping:

Learn AI Series (#105) - Monte Carlo Methods What will I learn You will learn how to estimate value functions from complete episodes of experience -- no model of the environment required; first-visit vs