EVENT

JFLI seminar on "Explainable sparse models" by Nataliya Sokolovska

We are pleased to inform you about the upcoming seminar by Nataliya Sokolovska, full professor at Sorbonne University (France), titled : "Explainable sparse models" Everyone interested is cordially invited to attend!

Title:

Explainable sparse models

Abstract:

Traditional machine learning methods consider features to be independent, even if they are highly correlated in a real task. However, in the last decade, an increasing attention was devoted to non-additive integrals, such as Choquet and Sugeno integrals, enabling the modeling of interactions among variables and providing a fine control of synergies between them. These integrals are more and more used in the context of supervised learning, i.e., where an algorithm has access to observations and their labels. The non-additive integrals are introduced into loss functions to learn reliable predictive models. We aim to take the best of two worlds, machine learning and decision theory, both actively developed in Artificial Intelligence, to propose adaptive and interpretable evaluation approaches and contribute to produce reliable predictive models.
Our main motivation is to explore non-additive capacities to construct compact interpretable models. In particular, we are interested to efficiently optimise an objective (loss) function which is based on a compact (sparse) Choquet integral.

Speaker Bio:

Nataliya Sokolovska, full professor at Sorbonne University (France)

Time/Date:

10:30-12:00 / Monday, December. 9 , 2024

Place:

Room 1810, NII

Contact:

If you would like to join, please contact by email.
Email : boudin[at]nii.ac.jp

Link:

Japanese-French Laboratory for Informatics