Not All Samples Are Created Equal: Deep Learning with Importance Sampling.

Angelos Katharopoulos François Fleuret

Published in: CoRR (2018)

Keyphrases

deep learning
importance sampling
monte carlo
markov chain
kalman filter
unsupervised learning
particle filter
machine learning
approximate inference
mental models
training set
markov chain monte carlo
graphical models
weakly supervised
belief propagation
particle filtering
message passing
feature selection