Master 2024/2025

Tensor Decompositions and Tensor Networks in Artificial Intelligence

Type: Elective course (Math of Machine Learning)

When: 1 year, 2, 3 module

Open to: students of one campus

Language: English

Course Syllabus

Abstract

This course covers major topics in tensor decompositions and tensor networks for modern applications in Machine learning (ML), SignalProcessing (SP), Deep Neural Networks (DNN), Multiviews, and Data Fusion.The goal of the course is to provide students with a background in mathematics, especially in linear and multilinear algebra, statistics,useful computational tools.The emphasis of this course is on "learning by experimenting and programming". Students will learn various tensor decomposition modelsand state-of-the-art algorithms for each model. Students will study also challenging problems, e.g., degeneracy, stability, modelselection, and know how to deal with them in practice, make the models scalable for big data, solve the complex tensor networks.Lectures will be strongly coupled with a number of hands-on sessions.