Associate Professor, Department of Computer Sciences, University of Wisconsin - Madison
3 papers 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.
