About

Olivier Verdier

I’m a consultant, in artificial intelligence, statistics, algorithms and mathematics in general.

I was previously associate professor of mathematics in the Information and Communication Technology group in the Department of Computer Science, Electrical Engineering and Mathematical Sciences of the Western Norway University of Applied Sciences.

Interests

My main interests are:

  • Large Language models and their applications in companies
  • Machine Learning, Deep Learning and Bayesian Statistics
  • Scientific Computing, Geometric Integration

Previous Research Themes

The main themes of my research revolved around the integration of differential equations (ODEs or PDEs) on manifolds.

There are four themes on which I particularly focused on:

  • Structure preservation How to preserve the differential invariant (volume, symplectic), first integrals (energy), fixed points, periodic orbits, of constrained systems?
  • Symmetries Manifolds are often endowed with a natural symmetry: how can integrators respect that symmetry? A particularly important case is that of affine symmetries, related to scaling invariance and the anisotropy problem.
  • Constraint reduction A combination of differential equations and algebraic equations leads to a reduction algorithm to obtain a differential equation on a manifold: does the reduction halt? how does the integrator preserve the underlying manifold?
  • Software development Development, packaging and distribution of scientific computing software.