OVERVIEW

The course will provide an overview of principles behind neural networks and deep architectures, providing an overview of classical and recent approaches

AIMS AND CONTENT

LEARNING OUTCOMES

Learning how to use advanced machine learning algorithms, including learning data representation (dictionaries and metric), deep learning, and learning in dynamic environment (online, active and reinforcement learning), by grasping the underlying computational and modeling issues.

AIMS AND LEARNING OUTCOMES

Students will be provided with an overview of neural networks and deep architectures, starting from basic principles to more advanced concepts. An overview of the different types of architecture will be presented. Also, recent methodologies will be introduced to discuss practical problems related for instance to computational aspects, data requirements, and generalization abilities.
Hands-on activities, in which students will practice the use of neural networks, will always complement the theoretical classes.  The students will deepen their capability of critically analysing the results.

PREREQUISITES

Basic of Machine Learning, programming (preferrable in Python)

TEACHING METHODS

Theoretical classes will be coupled with practical lab sessions
Students will be asked to work in groups during such lab sessions.

SYLLABUS/CONTENT

The course will cover the following topics:

• Neural Networks
• Convolutional Neural Networks
• Recurrent Neural Networks
• LSTMs
• Transformers
• Graph Neural Networks
• Autoencoders and GANs
• Deep clustering
• Representation Learning strategies
• Transfer Learning and domain adaptation

RECOMMENDED READING/BIBLIOGRAPHY

The material will provided by the instructors, see the course Aulaweb page for additional references.

TEACHERS AND EXAM BOARD

Exam Board

NICOLETTA NOCETI (President)

VITO PAOLO PASTORE

VITTORIO MURINO (President Substitute)

LESSONS

LESSONS START

In agreement with the academic calendar approved by the Committee of the Study Courses in Informatics and Computer Science

Class schedule

The timetable for this course is available here: Portale EasyAcademy

EXAMS

EXAM DESCRIPTION

The exam will consist in two main parts:

• a project (in Python) that will be presented in a short seminar (no project if number of credits < 9)
• an oral exam

ASSESSMENT METHODS

The exam will evaluate the overall understanding of the topics of the course, the capability to generalize the concepts to unseen problems and analyse the obtained results.
Clarity of exposition, completeness of the concepts, quality of the proposed solutions and critical thinking will be taken into account.

Exam schedule