Charles Hermite
PulseAugur coverage of Charles Hermite — every cluster mentioning Charles Hermite across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
New algorithms for learning positive ReLU networks under Gaussian inputs
Researchers have developed polynomial-time algorithms for learning positive ReLU networks, specifically addressing scenarios with standard Gaussian inputs. The algorithms achieve polynomial sample and real-arithmetic co…
-
Symmetry-Aware Feature Learning Shows Polynomial Advantage Over Agnostic Methods
Researchers have demonstrated a significant polynomial separation in sample complexity between symmetry-aware and symmetry-agnostic feature learning for multi-index models. The study, which focuses on high-dimensional G…
-
New theory explains neural network learning during loss plateaus
Researchers have developed a new theoretical framework for understanding how neural networks learn, particularly during periods where the overall loss plateaus. The study focuses on two-layer networks with ReLU and Leak…
-
New Adam Theorem Unveiled for Spectral Heavy-Tail Onset in ML Models
Researchers have developed a comprehensive theoretical framework, termed a "full Adam theorem," to analyze the spectral heavy-tail onset in Gaussian Stein-Hermite teacher-student models. This theorem meticulously detail…
-
New theory explains incremental learning in shallow neural networks
Researchers have developed a new theoretical framework for understanding incremental learning in shallow neural networks. This work focuses on polynomial-width two-layer networks trained on orthogonal multi-index target…