Reinforcement Learning for AI Agents Training in Thimphu, Bhutan

This course provides an in-depth introduction to Reinforcement Learning (RL) with a strong emphasis on designing intelligent, goal-driven AI agents. Participants will learn how agents learn optimal behavior through interaction with environments and rewards. The course blends theory, algorithms and hands-on programming using Python and RLlib. Ideal for professionals aiming to build agents for robotics, game AI, automation, or autonomous decision systems.

Foundations of Reinforcement Learning

• Overview of Machine Learning paradigms

• What is Reinforcement Learning?

• Markov Decision Processes (MDPs)

• Agents, environments, states, actions, rewards

• Exploration vs exploitation

Value-Based Methods

• Monte Carlo methods

• Temporal Difference (TD) Learning

• Q-Learning and SARSA algorithms

• Introduction to function approximation

Policy-Based and Deep Reinforcement Learning

• Policy Gradient methods

• REINFORCE algorithm

• Deep Q Networks (DQN)

• Challenges: Stability, overfitting, delayed rewards

Hands-on Exercises

Case Studies

Summary and Conclusion