Study of Deep Generative Models for Inorganic Chemical Compositions
This work addresses a gap for material researchers who lack crystal structure data, but it is incremental as it adapts existing generative models to a new domain.
The study tackled the problem of generating inorganic chemical compositions without crystal structure data, which is often unavailable, by comparing conditional VAE and GAN models, finding that CondGAN with a bag-of-atom representation and physical descriptors produced better compositions and evaluated methods like Metropolis-Hastings-based valency modification for material discovery.
Generative models based on generative adversarial networks (GANs) and variational autoencoders (VAEs) have been widely studied in the fields of image generation, speech generation, and drug discovery, but, only a few studies have focused on the generation of inorganic materials. Such studies use the crystal structures of materials, but material researchers rarely store this information. Thus, we generate chemical compositions without using crystal information. We use a conditional VAE (CondVAE) and a conditional GAN (CondGAN) and show that CondGAN using the bag-of-atom representation with physical descriptors generates better compositions than other generative models. Also, we evaluate the effectiveness of the Metropolis-Hastings-based atomic valency modification and the extrapolation performance, which is important to material discovery.