Contact

Bio

I am a full professor at Université Rennes 2, a position I have held since September 2022 after nine years as an assistant professor. I currently conduct my research at Inria Rennes within the MALT team. My work lies at the intersection of machine learning and time series, with a focus on optimal transport, differentiable programming, and time series foundation models. I also head the Master's program in Applied Mathematics and Statistics at Université Rennes 2 and teach at Institut Polytechnique de Paris (École Polytechnique).

News

• Vision Transformer Finetuning Benefits from Non-Smooth Components: Ambroise's paper has been accepted at ICML'26.
• Differentiable Generalized Sliced Wasserstein Plans: our paper has been accepted at NeurIPS'25.
• One for all and all for one: Efficient Computation of Partial Wasserstein Distances on the Line: our paper has been accepted at ICLR'25.
• Check out tslearn: a machine learning toolkit dedicated to time series data.

Students and Post-docs (past & present)