Day 3 - Reinforcement Learning Foundations

Day 3: Reinforcement Learning Foundations

Agents, environments and rewards, Markov decision processes, value functions and the Bellman equation, and Q-learning.

Topic	Description & Objectives	Resource
RL Foundations	The agent-environment loop, Markov decision processes, value functions, the Bellman equation and Q-learning (Labs 1 to 3)	Open
Lab 1: Introduction to Reinforcement Learning	Environments, agents, rewards and returns, and policies in Gymnasium	Notebook Colab
Lab 2: Writing Policies by Hand	Write rule-based policies for CartPole and FrozenLake and score them by average return, then two challenges: balance the pole for all 500 steps, and cross the slippery lake	Notebook Colab
Lab 3: Q-Learning	Build a Q-table, balance exploration and exploitation, and apply the Bellman update	Notebook Colab
Optional: Writing Policies by Hand (Solution)	Worked solution to Lab 2	Notebook Colab