Exploratory action

MEGAVOLT

Machine Learning and Evolution Equations
Machine Learning and Evolution Equations

A fierce scientific  problem is to build a mathematical  and numerical bridge between, on the one hand the mathematical modeling of complex phenomena usually undertaken with partial differential equations/ordinary differential equations/statistical differential equations, and on the other hand the computational statistical approach which relies on the exploitation of databases and which has made tremendous progres in the recent years.  One key feature is to understand the stability of learning with respect to the time evolution of a given phenomenon, but scarce is known. The MEGAVOLT project will explore the use  modern Machine Learning techniques applied to various non stationary in time differential equations and will seek to develop an operational  theory of such problems.

Internal partner(s)
Centre de Paris, Sorbonne Université

Contacts

Bruno Despres

Scientific leader