Parallelizing Training of Deep Generative Models on Massive Scientific Datasets.
Sam Ade JacobsJim GaffneyTom BensonPeter B. RobinsonJ. Luc PetersonBrian K. SpearsBrian Van EssenDavid HysomJae-Seung YeomTim MoonRushil AnirudhJayaraman J. ThiagarajanShusen LiuPeer-Timo BremerPublished in: CLUSTER (2019)
Keyphrases
- generative model
- deep belief networks
- restricted boltzmann machine
- probabilistic model
- discriminatively trained
- discriminative learning
- mixture model
- deep learning
- discriminative models
- hierarchical hidden markov models
- generative and discriminative models
- object categories
- conditional random fields
- prior knowledge
- training set
- object detection
- semi supervised
- em algorithm
- expectation maximization
- maximum entropy principle
- training data
- maximum likelihood
- graphical models
- image processing