Summary of Week 13 -- Capsule 2 -- A first RL Algorithm

00:00:00 - 00:15:00

This video covers the first visit Monte Carlo algorithm, a model-free reinforcement learning algorithm. The algorithm makes a few assumptions about the environment and uses a full episode to update the policy. This video covers how the algorithm works and how it can be used to calculate the value function of a policy.

• 00:00:00 In this video, we introduce a model-free algorithm for reinforcement learning, called monte carlo. The algorithm makes a few assumptions about the environment, and then uses a full episode to update the policy.
• 00:05:00 The video introduces a first RL algorithm, first visit Monte Carlo, which calculates the value function of a policy given an environment and a set of cards. This algorithm can be thought of as a sampling procedure, estimating the expected future sum of discounted rewards.
• 00:10:00 This video introduces a first-order RL algorithm for a grid world, demonstrating how it calculates the value of a state according to its rewards and punishments.
• 00:15:00 This video introduces the concept of algorithms for reinforcement learning, including model-based and model-free methods. It also covers the first visit Monte Carlo algorithm.