Not All Samples Are Created Equal: Deep Learning with Importance Sampling.
Angelos KatharopoulosFrançois FleuretPublished in: ICML (2018)
Keyphrases
- deep learning
- importance sampling
- monte carlo
- unsupervised learning
- markov chain
- machine learning
- kalman filter
- particle filter
- particle filtering
- weakly supervised
- data sets
- approximate inference
- supervised learning
- training set
- computer vision
- hidden markov models
- latent variables
- bayesian inference
- markov chain monte carlo