Variational policy search using sparse Gaussian process priors for learning multimodal optimal actions

Hikaru Sasaki, Takamitsu Matsubara