C
search
search
reviewers
reviewers
feeds
feeds
assignments
assignments
settings
logout
Gradient Descent with Identity Initialization Efficiently Learns Positive-Definite Linear Transformations by Deep Residual Networks.
Peter L. Bartlett
David P. Helmbold
Philip M. Long
Published in:
Neural Comput. (2019)
Keyphrases
</>
positive definite
linear transformations
linear transformation
covariance matrix
kernel function
parameter space
geometric structure
machine learning
objective function
image classification
loss function
kernel matrix
feature extraction
similarity measure
distance function
locally optimal