Program Overview

Program Overview

The Lecture series on scientific machine learning is a course that presents ongoing work on how scientific questions can be tackled using machine learning. Machine learning enables extracting knowledge from data computationally and in an automatized way.

Frequency

The course is offered every 2 years.

Summary

This lecture presents examples of how machine learning is influencing the very scientific method. It covers recent works in this direction, including the use of machine learning to address scientific questions in physics, chemistry, material science, and biology.

Content

The course will present some recent works in this direction. In the first part of the course, works of different EPFL laboratories that use machine learning to address scientific questions in physics, chemistry, material science, and biology will be presented. Professors involved include: Lenka Zdeborova, Giuseppe Carleo, Michele Ceriotti, Philippe Schwaller, Alexander Mathis, Anne-Florence Bitbol, David Harvey. Examples of problems covered include neural-network enhanced solutions of the Schrodinger equation in a variety of contexts, machine learning for the prediction and rationalization of chemical and physical properties of materials, analysis of proteins from their sequence and structure, or automated data analysis and modeling brain-function in neuroscience, application in astrophysics. In the second part of the lecture, students will read, present, and discuss selected recent articles on the subject.

Requirements

Prior basic notions of machine learning are required, any of the introductory courses to machine learning is suitable. This lecture is accessible to all students across disciplines interested in natural and computational sciences.

Learning Outcomes

By the end of the course, the student must be able to:

• orient themselves in scientific problems where tools of machine learning may be applied
• be able to understand and discuss the related literature
• identify suitable questions and tools

Programs

The course is part of the following programs:

• Physics, Doctoral School
  • Number of places: 50
  • Exam form: Oral presentation (session free)
  • Subject examined: Lecture series on scientific machine learning
  • Courses: 14 Hour(s)
  • Project: 28 Hour(s)
  • Type: optional
• Computational and Quantitative Biology, Doctoral School
  • Number of places: 50
  • Exam form: Oral presentation (session free)
  • Subject examined: Lecture series on scientific machine learning
  • Courses: 14 Hour(s)
  • Project: 28 Hour(s)
  • Type: optional
• Computer and Communication Sciences, Doctoral School
  • Number of places: 50
  • Exam form: Oral presentation (session free)
  • Subject examined: Lecture series on scientific machine learning
  • Courses: 14 Hour(s)
  • Project: 28 Hour(s)
  • Type: optional
• Electrical Engineering, Doctoral School
  • Number of places: 50
  • Exam form: Oral presentation (session free)
  • Subject examined: Lecture series on scientific machine learning
  • Courses: 14 Hour(s)
  • Project: 28 Hour(s)
  • Type: optional
• Materials Science and Engineering, Doctoral School
  • Number of places: 50
  • Exam form: Oral presentation (session free)
  • Subject examined: Lecture series on scientific machine learning
  • Courses: 14 Hour(s)
  • Project: 28 Hour(s)
  • Type: optional
• Neuroscience, Doctoral School
  • Number of places: 50
  • Exam form: Oral presentation (session free)
  • Subject examined: Lecture series on scientific machine learning
  • Courses: 14 Hour(s)
  • Project: 28 Hour(s)
  • Type: optional

Keywords

machine learning, neural networks, scientific machine learning

Teachers

The course is taught by:

• Bitbol Anne-Florence Raphaëlle
• Carleo Giuseppe
• Ceriotti Michele
• Harvey David Richard
• Mathis Alexander
• Schwaller Philippe
• Zdeborová Lenka

Language

The course is taught in English.

Credits

The course is worth 2 credits.