Factorized Gaussian Process Variational Autoencoders
This addresses scalability issues in structured latent variable modeling for researchers and practitioners working with datasets containing independent features.
The paper tackles the cubic inference time complexity of Gaussian process variational autoencoders by factorizing the latent kernel across independent auxiliary features, achieving significant speed-up while maintaining comparable performance to non-scalable approaches.
Variational autoencoders often assume isotropic Gaussian priors and mean-field posteriors, hence do not exploit structure in scenarios where we may expect similarity or consistency across latent variables. Gaussian process variational autoencoders alleviate this problem through the use of a latent Gaussian process, but lead to a cubic inference time complexity. We propose a more scalable extension of these models by leveraging the independence of the auxiliary features, which is present in many datasets. Our model factorizes the latent kernel across these features in different dimensions, leading to a significant speed-up (in theory and practice), while empirically performing comparably to existing non-scalable approaches. Moreover, our approach allows for additional modeling of global latent information and for more general extrapolation to unseen input combinations.