OVERVIEW

What is a machine learning algorithm? Why is machine learning playing a primary role in physics? Which problems can be optimized using it? What is the most suitable algorithm to solve my physics problem?

These are some of the questions that this course aims to answer, providing students with the state-of-the-art knowledge regarding the usage and understanding of artificial intelligence algorithms applied to physics. The course also focuses on developing a critical comprehension of results, exploring the development of future algorithms, and the most promising technologies.

AIMS AND CONTENT

LEARNING OUTCOMES

Teaching aims to develop skills in understanding and implementing machine learning algorithms applied to physics problems. Major neural algorithms will be described and discussed with various practical examples of how they are used to solve physics problems.

AIMS AND LEARNING OUTCOMES

The course aims to provide the conceptual, theoretical, and methodological tools for a clear understanding of machine learning algorithms used in physics. To achieve this goal, the necessary techniques to comprehend the most modern neural networks and their applications to physics problems will be described and put into practice.

By the end of the course, students will be able to:

1. Understand the functioning of feed-forward, convolutional, and recurrent neural networks.
2. Utilize these networks to build more advanced architectures, such as deep-sets, graph neural networks and transformers.
3. Critically analyze scientific papers related to the usage of these techniques for solving physics problems.
4. Independently set up and implement every step for developing a machine learning algorithm that is optimal for their specific problem.

PREREQUISITES

The course is self-contained. General knowledge of particle physics is useful but not necessary. Similarly, knowledge of the Python programming language is helpful (knowledge acquired during the undergraduate studies) but not essential.

TEACHING METHODS

The course has a strong practical component. The theory of machine learning algorithms will be discussed in lectures, and then they will be implemented in dedicated practical sessions. This approach allows students to gain a solid understanding of the concepts and then apply them directly through hands-on implementation, ensuring a comprehensive learning experience.

SYLLABUS/CONTENT

The course aims to:

• Introduce the concepts of minimization algorithms for a scalar functional (the loss function).
• Provide the necessary tools for practical course execution, such as Python, Tensorflow, and Pytorch.
• Cover dense neural networks and examples of their applications in physics.
• Explore convolutional neural networks and examples of their applications in physics.
• Discuss recurrent neural networks and examples of their applications in physics.
• Investigate graph neural networks: inductive bias and examples of their applications in physics.
• Examine attention mechanisms: transformers and examples of their applications in physics.
• Study generative neural networks and examples of their applications in physics.
• Provide an overview of differentiable programming.

The course encompasses these topics to provide students with a comprehensive understanding of machine learning algorithms in the context of physics applications.

TEACHERS AND EXAM BOARD

Exam Board

FRANCESCO ARMANDO DI BELLO (President)

ANDREA COCCARO

RICCARDO TORRE (Substitute)

LESSONS

LESSONS START

Please refer to the calendar at the following link: https://corsi.unige.it/corsi/9012/studenti-orario

Class schedule

The timetable for this course is available here: Portale EasyAcademy

EXAMS

EXAM DESCRIPTION

The exam consists of a written part where students will be asked to solve an exercise and an oral part to assess their understanding of the course material. The written section will test the practical application of concepts, while the oral part will evaluate their theoretical knowledge.

ASSESSMENT METHODS

The committee will evaluate both the written work and the oral discussion. Both components, the written elaboration and the oral presentation, will be taken into consideration during the assessment process.

FURTHER INFORMATION

Students who have valid certification of physical or learning disabilities on file with the University and who wish to discuss possible accommodations or other circumstances regarding lectures, coursework and exams, should speak both with the instructor and with Professor Sergio Di Domizio (sergio.didomizio@unige.it), the Department’s disability liaison.