Complexity-aware loss function for fast neural networks with early exits.

Lev Teplyakov Sergey Gladilin Evgeny Shvets

Published in: ICMV (2019)

Keyphrases

loss function
neural network
pairwise
logistic regression
learning to rank
support vector
empirical risk
boosting algorithms
risk minimization
hinge loss
boosting framework
objective function
worst case
back propagation
convex loss functions
square loss
computational complexity
machine learning