|
|
| Academic Year |
2023/2024 |
| Name |
Dott. - MI (1380) Ingegneria dell'Informazione / Information Technology |
| Programme Year |
1 |
| ID Code |
061642 |
| Course Title |
REINFORCEMENT LEARNING |
| Course Type |
MONO-DISCIPLINARY COURSE |
| Credits (CFU / ECTS) |
5.0 |
| Course Description |
Reinforcement learning deals with solving sequential decision-making problems when no (or minimal) prior information is available. Solving sequential decision-making problems means finding their optimal control policies. Using reinforcement-learning algorithms, the optimal policy is learned through the direct interaction between the agent (or controller) and the system to be controlled. The course will introduce the main modeling frameworks, analyze the most relevant reinforcement-learning techniques, and, finally, show some interesting applications of these techniques to real-world domains.
1) Models
1.1) Finite Markov Decision Processes
1.2) Continuous Markov Decision Processes
1.3) Markov Games
2) Algorithms
2.1) Dynamic Programming (Policy Iteration, Value Iteration)
2.2) Model-free algorithms for prediction (Monte Carlo, Temporal Difference)
2.3) Model-free algorithms for control (Q-learning, SARSA)
2.4) Online RL with value function approximation (GTD, TDC)
2.5) Batch RL with value function approximation (LSTD, LSPI, FQI)
2.6) Policy Search (policy gradient, natural policy gradient, PGPE)
2.7) Deep Reinforcement Learning (DQN, TRPO, PPO, DDPG, SAC)
2.8) Planning and Learning (MCTS, AlphaZero)
2.9) Multi-agent Reinforcement Learning (overview)
3) Applications
3.1) Autonomous Driving
3.2) Water Resources Management
3.3) Algorithmic Trading |
| Scientific-Disciplinary Sector (SSD)
|
|
SSD Code
|
SSD Description
|
CFU
|
|
ING-INF/05
|
INFORMATION PROCESSING SYSTEMS
|
5.0
|
|
|
Alphabetical group
|
Name
|
Teaching Assignment Details
|
|
From (included)
|
To (excluded)
|
|
A
|
ZZZZ
|
Restelli Marcello, Metelli Alberto Maria
|
|
|
