C9789 - ARTIFICIAL INTELLIGENCE AND DECISION MAKING

Learning outcomes

By the end of this course, students will be able to: understand and apply key concepts from dynamic programming and optimization to AI decision problems; model and analyze decision-making processes in dynamic and uncertain environments; formulate and solve multi-objective and multi-agent decision problems; understand the role of dynamic games in the design of intelligent agents; explore applications of these principles in reinforcement learning, federated learning, and adversarial AI systems.

Course contents

The course provides a unified treatment of decision making techniques and their use in modern artificial intelligence. The contents are organized as follows:

1. Static optimization
2. Dynamic programming
3. The Bellman equation
4. Decision making under uncertainty
5. Multi-objective optimization
6. Dynamic and differential games
7. Supervised and unsupervised learning
8. Reinforcement learning
9. Multi-agent learning
10. Federated and adversarial learning

Readings/Bibliography

Lecture notes and slides provided by the instructor.

H. Kunze, D. La Torre, A. Riccoboni, and M. Ruiz Galan (Eds.), "Engineering Mathematics and Artificial Intelligence: Foundations, Methods, and Applications," CRC Press (Taylor & Francis), 2023. (Series: Mathematics and Its Applications.)

S. Ben-David, G. Curigliano, D. Koff, B. A. Jereczek-Fossa, D. La Torre, and G. Pravettoni (Eds.), "Artificial Intelligence for Medicine: An Applied Reference for Methods and Applications," Elsevier (Academic Press), 2024.

M. Corazza, R. Garcia, F. S. Khan, D. La Torre, and H. Masri (Eds.), "Artificial Intelligence and Beyond for Finance," World Scientific, 2024. (Series: Transformations in Banking, Finance and Regulation, Vol. 15.)

F. P. Appio, D. La Torre, F. Lazzeri, H. Masri, and F. Schiavone (Eds.), "Impact of Artificial Intelligence in Business and Society: Opportunities and Challenges," Routledge (Taylor & Francis), 2024. (Series: Routledge Studies in Innovation, Organizations and Technology.)

S. Boyd and L. Vandenberghe, "Convex Optimization," Cambridge University Press, 2004.

R. S. Sutton and A. G. Barto, "Reinforcement Learning: An Introduction," 2nd ed., MIT Press, 2018. (Series: Adaptive Computation and Machine Learning.)

Teaching methods

Theory, exercises, tutorials, class work, project.

Assessment methods

For attending students: project presentation.

For non-attending students: oral examination.

Teaching tools

Generative Artificial Intelligence.

Office hours

See the website of Davide La Torre