MI is All You Need: Understanding Complex Multivariate Systems Through the Lenses of GenAI

-

B1 L4 R4102

This talk explores the use of score functions and denoising neural networks for efficient estimation of mutual information and novel extensions to information theory that describe complex multivariate interactions, demonstrating applications in automotive sensing, semantic alignment, and neuroscience.

In this talk we will see how to leverage score functions to estimate mutual information, as well as novel extensions to information theory that aim at describing unique, redundant and synergistic interactions among arbitrary random variables composing complex, multivariate information systems.

We will briefly cover the theory that makes it possible to transform such difficult problems into efficient, amortized learning of denoising neural networks, and present their applications to a variety of contexts, including virtual sensing for automotive, semantic alignment, and neuroscience

Presenters

Pietro Michiardi, Professor, Department Head, Data Science Department, EURECOM

Brief Biography

Pietro Michiardi received his M.S. in Electrical Engineering from Politecnico di Torino and his  M.S. in Computer Science from EURECOM. Pietro received his Ph.D. in Computer Science from Telecom ParisTech, and his HDR (Habilitation) from UCA.

Today, Pietro is a Full Professor and head of the Data Science Department at EURECOM. In his work, Pietro focuses on scalable machine learning algorithms, Bayesian modelling and sampling, generative modelling and representation learning. Pietro is interested in developing a theoretical understanding of methods that learn from data, in methodological aspects of computationally efficient inference approaches, and their application to various domains such as the automotive and telecommunication industries, as well as neuroscience.

In the past, Pietro worked on a wide range of research topics, including: computer networks and their performance evaluation, applied cryptography, applied game theory, distributed systems, and data management systems