On Learnability via Gradient Method for Two-Layer ReLU Neural Networks in Teacher-Student Setting.

Shunta Akiyama Taiji Suzuki

Published in: ICML (2021)

Keyphrases

gradient method
neural network
teacher student
convergence rate
step size
optimization methods
negative matrix factorization
professional development
simulated annealing
learning algorithm
genetic algorithm
training data
digital libraries
training set
online learning
sensemaking
teaching strategies
data mining