Two Sides of One Coin: the Limits of Untuned SGD and the Power of Adaptive Methods.

Junchi Yang Xiang Li Ilyas Fatkhullin Niao He

Published in: CoRR (2023)

Keyphrases

significant improvement
similarity measure
benchmark datasets
neural network
pairwise
machine learning methods
feature selection
image processing
online learning
empirical studies
step size