Reinforcement Learning

Reinforcement learning is one the main paradigms of modern machine learning, artificial intelligence and robotics, with wide applications for decision-making and for the training of autonomous agents. This course

provides an understanding of the foundation of reinforcement learning, analyzes classical reinforcement

learning algorithms, and shows how practical problems can be modelled and solved with a reinforcement

learning approach.

Knowledge

The student should be able to:

• Understand the foundational concepts of reinforcement learning;
• Formalize and express problems in the reinforcement learning framework;
• Appreciate differences, strengths and limitations of classical reinforcement learning algorithms;
• Evaluate and decide how to solve reinforcement learning problems.

Skills

The student should be able to:

• Solve reinforcement learning problems using classical algorithms;
• Implement reinforcement learning algorithms and adapt them to specific problems.

General competence

The student should be able to:

• Understand when and how specific problems may be successfully solved through the reinforcement learning paradigm;
• Understand and critically evaluate state-of-the-art algorithms