Timezone: »

Scalable Inference for Gaussian Process Models with Black-Box Likelihoods
Amir Dezfouli · Edwin Bonilla

Mon Dec 07 04:00 PM -- 08:59 PM (PST) @ 210 C #33

We propose a sparse method for scalable automated variational inference (AVI) in a large class of models with Gaussian process (GP) priors, multiple latent functions, multiple outputs and non-linear likelihoods. Our approach maintains the statistical efficiency property of the original AVI method, requiring only expectations over univariate Gaussian distributions to approximate the posterior with a mixture of Gaussians. Experiments on small datasets for various problems including regression, classification, Log Gaussian Cox processes, and warped GPs show that our method can perform as well as the full method under high levels of sparsity. On larger experiments using the MNIST and the SARCOS datasets we show that our method can provide superior performance to previously published scalable approaches that have been handcrafted to specific likelihood models.

Author Information

Amir Dezfouli (The University of New South Wales)
Edwin Bonilla (University of New South Wales)

More from the Same Authors