Maxence Noble

Hello ! I am a third-year PhD candidate in Machine Learning at Centre de Mathématiques Appliquées (CMAP) at École Polytechnique, advised by Alain Durmus. Prior to this, I graduated with a MSc. degree in Applied Mathematics from École Polytechnique (“Cycle Ingénieur”) and a MRes. degree in Mathematics, Vision and Learning (“MVA”) from École Normale Supérieure Paris-Saclay.
Current status. Since mid-March 2025, I have been interning at Jasper AI, in the French research team, to work on speeding up the inference of diffusion models for large-scale image restoration tasks. I will be returning to my PhD in mid-September 2025.
Research interests. I am interested in the study of problems at the intersection of generative modelling, sampling, dynamic optimal transport and stochastic optimal control (with a special focus on the Schrödinger Bridge problem and score-based diffusion models).
Environmental awareness. Similarly to many researchers in machine learning, I feel concerned by the environmental impact of our research field, especially about the unsustainable circumstances of centralized worldwide conferences (which are getting bigger and bigger). This has motivated me to join the Neurips@Paris initiative. Feel free to reach me out about this topic !
news
Mar 06, 2025 | My reflexion paper Improving the evaluation of samplers on multi-modal targets co-authored with Louis Grenioux and Marylou Gabrié has been accepted at the workshop Frontiers in Probabilistic Inference : Learning meets Sampling at ICLR 205 ! I am very excited to attend to this event (probably, the workshop most closely aligned with the content of my PhD !) : see you in Singapore in late April ! |
---|---|
Jan 23, 2025 | In the next days, I will be presenting the results from Learned Reference-based Diffusion Sampling for multi-modal distributions at Google Deepmind’s workshop on diffusion models as well as at the Center for Data Science’s reading group (ENS Ulm). |
Jan 20, 2025 | My latest paper Learned Reference-based Diffusion Sampling for multi-modal distributions co-authored with Louis Grenioux, Marylou Gabrié and Alain Durmus has been accepted at ICLR 205 ! |
Sep 01, 2024 | Along with some talented French researchers, I am co-organizing the 4-th edition of NeurIPS@Paris, which will take place in December ! |
Jun 01, 2024 | I gave a talk on Riemannian constrained sampling at the 4-th Italian Meeting on Probability and Mathematical Statistics (Rome, Italy). |