81940 - Artificial Intelligence

Learning outcomes

At the end of the course the student will know the principal languages, the modeling techniques and the reasoning methods that are at the base of artificial intelligence. In particular the student will be able to construct systems that exhibit intelligent behaviours, often simulating the behavior of human experts of a specific discipline. Moreover she will be able to model and solve simple constraint and optimization problems by using constraint programming.

Course contents

Introduction to artificial intelligence.
The principal technologies and applications of artificial intelligence.
The notion of agent.
Non informed search strategies.
Informed search strategies.
Search with adversaries.
Modeling of problems with costraints and CSP: basic notions.
Logic programming.
Constraint programming, basic notions of MiniZinc.
Introduction to machine learning.

Sub-symbolic computation and neural networks .

Large language models
Philosophical aspects and future challenges.

Readings/Bibliography

Russell, Norvig. Artificial Intelligence: A Modern Approach, 3rd Edition. Pearson (Intl) 2010 (US edition) and 2016 (Global edition).

Handouts provided during the course.

Teaching methods

Frontal lectures.

Assessment methods

The exam consists of the development of a project and its presentation and discussion.

Each academic year includes six exam sessions (“appelli”): two in the first session (January/February), three in the second (June/July), and one in the third (September). The dates of the sessions are published on Almaesami and may be subject to change. It is recommended to check 24 hours before the written exam to ensure that the date has not been modified.

To take part in an exam session, students must register on Almaesami no later than 7 days before the exam date. Students who fail to register on time will not be allowed to sit the exam and will have to wait for the next session.

The assessment consists of the presentation of the project and possible questions on the course topics.
The student must therefore demonstrate that they have actively worked on the project (which is carried out in groups) and that they have mastered the main AI techniques covered in the course.

The exam grade is determined according to the following criteria:

- if the project content, its presentation, and discussion are limited, with limited mastery of the course topics, the grade will range between 18–21;
- if the project content, its presentation, and discussion are satisfactory, with sufficient mastery of the course topics, the grade will range between 22–25;
- if the project content, its presentation, and discussion are good, with good mastery of the course topics, the grade will range between 26–29;
- if the project content, its presentation, and discussion are excellent, with full mastery of the course topics, the grade will be 30 or 30 with honors (“30 e lode”).

At the end of the oral exam, the student will be informed of the grade obtained and may decide whether to accept or refuse it. According to the university’s academic regulations, the possibility to refuse the grade must be granted by the instructor at least once for each course.

This course is integrated with course 28796 – COMPLEMENTI DI BASI DI DATI.
Consequently, the final grade can only be recorded within 3 days of obtaining both grades.

Teaching tools

We will use slides and specific software tools for AI applications

Office hours

See the website of Maurizio Gabbrielli

See the website of Stefano Pio Zingaro