Resume
Academic Achievements
- Paper 'DDM4IP' accepted at ICCV 2025; presented workshop paper on finetuning foundation models for molecular dynamics using kernels; paper on efficient Koopman operator learning with Nyström approximation available on ArXiv; completed PhD thesis titled 'Large Scale Kernel Methods for Fun and Profit'.
Research Experience
- Currently a PostDoc at INRIA Grenoble in the Thoth team, supervised by Julien Mairal. Presented K-Planes algorithm at CVPR and research on efficient Koopman operator learning at NeurIPS.
Education
- Completed PhD at the University of Genova, supervised by Lorenzo Rosasco.
Background
- Research interests include inverse problems for imaging, developing efficient algorithms for shallow learning, and applying kernel methods and shallow learning algorithms to scientific applications.
Miscellany
- Contact email: giacomo [dot] meanti [at] gmail [dot] com