PhD Student
CentraleSupelec & Inria & Mila

Alexandre Duval


I am a PhD Student at CentraleSupelec and Inria Saclay under the supervision of Prof. Fragkiskos Malliaros. From the start of 2022, I have also been attached to MILA (Quebec Artificial Intelligence Insitute), under the supervision of Prof. David Rolnick and Prof. Yoshua Bengio, as part of the group “AI for Climate Change”.

My research spans the broad field of Graph Machine Learning (Graph ML), i.e. an area of ML which focuses on non-euclidean data (e.g. graphs). After several projects on the explicability of Graph Neural Networks (GNN), graph pooling algorithms and graph-based text summarisation, I bifurcated towards projects using Graph ML to accelerate scientific discovery. In particular, I am working on electrocatalyst design to improve the energy efficiency of hydrogen storage, cement production and other energy-intensive industrial processes. This ambitious project first involves constructing scalable and expressive symmetry-preserving GNN architectures for materials modelling, before combining them with generative methods (e.g. GFlowNets) to both propose and evaluate new catalyst materials in an end-to-end pipeline.

In my spare time, I love playing football as well as a wide range of other outdoor activities. I am also a genuine art enthousiast, including painting, cinema and music.