Alexia Jolicoeur-Martineau
Alexia Jolicoeur-Martineau is a Senior Researcher and Team Lead at Samsung SAIT (Samsung Advanced Institute of Technology) AI Lab in Montreal. She holds a PhD in biostatistics from the Université de Montréal and has nine years of AI research experience spanning generative models, GANs, diffusion models, and reasoning architectures. She is best known for the Tiny Recursive Models (TRM) paper — "Less is More: Recursive Reasoning with Tiny Networks" — which simplified the 27M-parameter Hierarchical Recursive Models (HRM) into a single 7M-parameter network while improving benchmark performance on ARC-AGI-1 and winning 1st Place Paper Award at ARC Prize 2025. She is also widely cited for the Relativistic GAN (2018), trained overnight on a single GPU, which stabilized GAN training and accumulated over 1,000 citations.
“She figures out that you actually can back prop through all the way to the deep recursion, which actually improves performance much, much more. She makes the model three, four times smaller. Because it has that recursion it actually outperforms.”
Source→“She makes the model three, four times smaller. Because it has that recursion it actually outperforms.”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.