Thomas is a third year PhD student in theoretical chemistry under the supervision of Manuel Corral-Valero and Pascal Raybaud at IFPEN and Gabriel Stoltz and Tony Lelièvre at INRIA.

His thesis work mixes a rare events sampling method (Adaptive Multi-level Splitting, AMS) and machine learning to identify reaction coordinates. The rare events at stake are molecular dynamics transition between identified meta-stable states (reactant/products). The target is to use machine learning method to define a one dimensional reaction coordinate which will serve to sample reactive trajectories and compute their occurrence probability. The transition mechanism can then be studied by analyzing these trajectories and from their occurrence probability, the reaction rate constant can be computed.

Relevant topics include:

• Rare event sampling

• Machine learned collective variables and reaction coordinates

• Ab-initio molecular dynamics

• Machine Learning force fields

• Heterogeneous catalysis