GraphVAE: Towards Generation of Small Graphs Using Variational Autoencoders
This work addresses the lack of generative models for graphs, which is a problem for researchers in fields like chemistry and drug discovery, though it appears incremental as it adapts existing VAE techniques to a new domain.
The authors tackled the problem of generating small graphs, specifically molecules, by proposing GraphVAE, a variational autoencoder that outputs probabilistic fully-connected graphs directly, achieving results on the challenging task of molecule generation.
Deep learning on graphs has become a popular research topic with many applications. However, past work has concentrated on learning graph embedding tasks, which is in contrast with advances in generative models for images and text. Is it possible to transfer this progress to the domain of graphs? We propose to sidestep hurdles associated with linearization of such discrete structures by having a decoder output a probabilistic fully-connected graph of a predefined maximum size directly at once. Our method is formulated as a variational autoencoder. We evaluate on the challenging task of molecule generation.