Auxiliary Guided Autoregressive Variational Autoencoders
This addresses the challenge of preventing degenerate models in hybrid generative approaches for machine learning, though it appears incremental as it builds on existing methods with a novel training technique.
The paper tackles the problem of generative modeling of high-dimensional data by proposing a training procedure with an auxiliary loss function to control information capture in hybrid latent variable and autoregressive models, achieving state-of-the-art quantitative performance among such models and generating convincing samples.
Generative modeling of high-dimensional data is a key problem in machine learning. Successful approaches include latent variable models and autoregressive models. The complementary strengths of these approaches, to model global and local image statistics respectively, suggest hybrid models that encode global image structure into latent variables while autoregressively modeling low level detail. Previous approaches to such hybrid models restrict the capacity of the autoregressive decoder to prevent degenerate models that ignore the latent variables and only rely on autoregressive modeling. Our contribution is a training procedure relying on an auxiliary loss function that controls which information is captured by the latent variables and what is left to the autoregressive decoder. Our approach can leverage arbitrarily powerful autoregressive decoders, achieves state-of-the art quantitative performance among models with latent variables, and generates qualitatively convincing samples.