Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
Deep Learning as Neural Low-Degree Filtering: A Spectral Theory of Hierarchical Feature Learning
Yatin Dandi, Matteo Vilucchio, Luca Arnaboldi +2
Understanding how deep neural networks learn useful internal representations from data remains a central open problem in the theory of deep learning. We introduce Neural Low-Degree…
cs.LG2021
Genealogical Population-Based Training for Hyperparameter Optimization
Antoine Scardigli, Paul Fournier, Matteo Vilucchio +1
HyperParameter Optimization (HPO) aims at finding the best HyperParameters (HPs) of learning models, such as neural networks, in the fastest and most efficient way possible. Most r…