Assistant Professor, University of Warwick
1 paper at NeurIPS 2025
We provide a precise description on how the test risk scales with a suitable norm-based capacity when compared to a classical metric of model size.
