Timezone: »

 
VAEs meet Diffusion Models: Efficient and High-Fidelity Generation
Kushagra Pandey · Avideep Mukherjee · Piyush Rai · Abhishek Kumar
Event URL: https://openreview.net/forum?id=-J8dM4ed_92 »

Diffusion Probabilistic models have been shown to generate state-of-the-art results on several competitive image synthesis benchmarks but lack a low-dimensional, interpretable latent space, and are slow at generation. On the other hand, Variational Autoencoders (VAEs) have access to a low-dimensional latent space but, despite recent advances, exhibit poor sample quality. We present VAEDM, a novel generative framework for \textit{refining} VAE generated samples using diffusion models while also presenting a novel conditional forward process parameterization for diffusion models. We show that the resulting parameterization can improve upon the unconditional diffusion model in terms of sampling efficiency during inference while also equipping diffusion models with the low-dimensional VAE inferred latent code. Furthermore, we show that the proposed model exhibits out-of-the-box capabilities for downstream tasks like image superresolution and denoising.

Author Information

Kushagra Pandey (Indian Institute of Technology, Kanpur)

I am a graduate student in the Computer Science Department at IIT Kanpur. My current research interests are scalable bayesian inference and probabilistic deep generative models and their downstream applications to computer vision and computational genomics.

Avideep Mukherjee (Indian Institute of Technology Kanpur)
Piyush Rai (IIT Kanpur)
Abhishek Kumar (Google)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors