Variational Autoencoders with Riemannian Brownian Motion Priors.
Dimitrios KalatzisDavid EklundGeorgios ArvanitidisSøren HaubergPublished in: ICML (2020)
Keyphrases
- brownian motion
- stochastic process
- differential equations
- heavy tailed
- optimal control
- diffusion process
- denoising
- poisson process
- stochastic processes
- vector valued
- heavy traffic
- queue length
- closed form solutions
- markov chain
- bayesian framework
- image segmentation
- shape analysis
- anisotropic diffusion
- stochastic differential equations
- euclidean space
- optical flow
- prior distribution
- prior model
- prior knowledge
- multiscale
- arrival rate
- control strategy
- cost function
- inventory level
- reinforcement learning
- bayesian networks