A survey on Variational Autoencoders from a GreenAI perspective

TL;DR

This survey reviews recent Variational Autoencoder models emphasizing their architectural innovations and evaluates their energetic efficiency to promote sustainable AI practices.

Contribution

It provides a comprehensive comparison of recent VAE variations with a focus on energy efficiency and environmental impact, including mathematical details and implementation insights.

Findings

Some VAE models significantly reduce energy consumption.

Architectural choices impact the models' efficiency and quality.

The survey highlights the importance of Green AI in generative modeling.

Abstract

Variational AutoEncoders (VAEs) are powerful generative models that merge elements from statistics and information theory with the flexibility offered by deep neural networks to efficiently solve the generation problem for high dimensional data. The key insight of VAEs is to learn the latent distribution of data in such a way that new meaningful samples can be generated from it. This approach led to tremendous research and variations in the architectural design of VAEs, nourishing the recent field of research known as unsupervised representation learning. In this article, we provide a comparative evaluation of some of the most successful, recent variations of VAEs. We particularly focus the analysis on the energetic efficiency of the different models, in the spirit of the so called Green AI, aiming both to reduce the carbon footprint and the financial cost of generative techniques. For each architecture we provide its mathematical formulation, the ideas underlying its design, a detailed model description, a running implementation and quantitative results.