CAS logo CASRL SERIES

◆ A CAS LEARNING PATH · SELF-PACED · BROWSER-NATIVE

The Reinforcement Learning Series

A growing collection of interactive lectures that teach how agents master sequential decisions from trial and error — through live mazes, reproducible bandits, animated loops, and intuition before heavy math.

◈ THE ROAD AHEAD

One series, one path — from trial & error to optimal agents

Lecture 01 builds crystal-clear trial-and-error intuition first. Subsequent modules add formal Markov Decision Processes, value estimation, temporal-difference control, and policy optimization.

Intuition before math

Everyday analogies — learning to ride a bicycle, navigating a maze, choosing restaurants, pulling slot levers — ground every Reinforcement Learning term before equations appear.

Live interactive labs

Steer a 5x5 Robot Maze, animate the Agent-Environment loop one step at a time, sweep the Discount Factor slider, and benchmark Bandit strategies deterministically in the browser.

Zero setup · Instant visual intuition

Clean one-look reminders, expressive Agent scenes, mastery diagnostics, Light & Dark textbook modes, and 100% browser-native delivery.