Grounded representations through deep variational inference and dynamic programming.
Juan Sebastian OlierEmilia I. BarakovaMatthias RauterbergLucio MarcenaroCarlo S. RegazzoniPublished in: ICDL-EPIROB (2017)
Keyphrases
- variational inference
- dynamic programming
- bayesian inference
- topic models
- gaussian process
- probabilistic model
- variational methods
- posterior distribution
- mixture model
- probabilistic graphical models
- latent dirichlet allocation
- closed form
- exponential family
- graphical models
- exact inference
- approximate inference
- factor graphs
- hyperparameters
- semi supervised
- belief propagation
- information extraction