On Signal-to-Noise Ratio Issues in Variational Inference for Deep Gaussian Processes.
Tim G. J. RudnerOscar KeyYarin GalTom RainforthPublished in: ICML (2021)
Keyphrases
- signal to noise ratio
- gaussian processes
- gaussian process
- variational inference
- noise reduction
- regression model
- bayesian inference
- hyperparameters
- approximate inference
- multi task learning
- latent variables
- bayesian framework
- semi supervised
- model selection
- latent dirichlet allocation
- probabilistic graphical models
- posterior distribution
- conditional random fields
- edge detection