Predicting the Generalization Gap in Deep Networks with Margin Distributions.

Yiding Jiang Dilip Krishnan Hossein Mobahi Samy Bengio

Published in: ICLR (Poster) (2019)

Keyphrases

power law
heavy tailed distributions
complex networks
network structure
support vector
social networks
training set
probability distribution
neural network
network size
exponential distributions
complex systems
computer networks
joint distribution
network design
training error
genetic algorithm