Efficient training for future video generation based on hierarchical disentangled representation of latent variables.
Naoya FushishitaAntonio Tejero-de-PablosYusuke MukutaTatsuya HaradaPublished in: CoRR (2021)
Keyphrases
- latent variables
- hierarchical model
- structural svms
- latent space
- probabilistic model
- posterior distribution
- random variables
- video data
- hidden variables
- structured prediction
- variational bayes
- prior knowledge
- video content
- max margin
- topic models
- video sequences
- latent variable models
- supervised learning
- active learning
- semi supervised
- elastic net
- bayesian networks