Conditional Inference in Pre-trained Variational Autoencoders via Cross-coding
This addresses a problem for researchers and practitioners using VAEs who need efficient conditional inference for arbitrary queries, representing an incremental improvement over existing sampling methods.
The paper tackles the challenge of conditional inference in pre-trained Variational Autoencoders (VAEs) by proposing cross-coding to approximate latent variable distributions after conditioning on evidence, enabling query sample generation without retraining. The result shows that cross-coding variations outperform Hamiltonian Monte Carlo in quantitative and qualitative evaluations.
Variational Autoencoders (VAEs) are a popular generative model, but one in which conditional inference can be challenging. If the decomposition into query and evidence variables is fixed, conditional VAEs provide an attractive solution. To support arbitrary queries, one is generally reduced to Markov Chain Monte Carlo sampling methods that can suffer from long mixing times. In this paper, we propose an idea we term cross-coding to approximate the distribution over the latent variables after conditioning on an evidence assignment to some subset of the variables. This allows generating query samples without retraining the full VAE. We experimentally evaluate three variations of cross-coding showing that (i) they can be quickly optimized for different decompositions of evidence and query and (ii) they quantitatively and qualitatively outperform Hamiltonian Monte Carlo.