Gradient Descent with Identity Initialization Efficiently Learns Positive-Definite Linear Transformations by Deep Residual Networks - researchr publication authors

researchr

You are not signed in
Sign in
Sign up

Peter L. Bartlett, David P. Helmbold, Philip M. Long. Gradient Descent with Identity Initialization Efficiently Learns Positive-Definite Linear Transformations by Deep Residual Networks. Neural Computation, 31(3), 2019. [doi]

This author has not been identified. Look up 'Peter L. Bartlett' in GoogleThis author has not been identified. Look up 'David P. Helmbold' in GoogleThis author has not been identified. Look up 'Philip M. Long' in Google

runs on WebDSL