The Tackling the Underspecified (TAU) project was founded in 2017, gathering members of the former
TAO project (2003-2016) and building upon their expertise. Our research focuses on AI that is
- in safe interaction with our world ;
- empowered by scalable learning abilities ;
- tightly interacting with Physics, and in particular Statistical Physics, both at the methodological fundamental and at the application levels ;
- able to take advantage of, and participate in, the progress of human knowledge through selected machine learning applications, specifically ones modeled through Partial Differential Equations and simulators;
- created and manifested via competitive/cooperative international challenges.
