Physics guided deep learning generative models for crystal materials discovery
This work addresses the challenge of generating stable and synthesizable crystal materials for materials science, representing an incremental improvement over previous models.
The authors tackled the problem of generating physically feasible crystal structures using deep learning generative models, achieving higher physical feasibility and expanding from only cubic structures to more diverse forms.
Deep learning based generative models such as deepfake have been able to generate amazing images and videos. However, these models may need significant transformation when applied to generate crystal materials structures in which the building blocks, the physical atoms are very different from the pixels. Naively transferred generative models tend to generate a large portion of physically infeasible crystal structures that are not stable or synthesizable. Herein we show that by exploiting and adding physically oriented data augmentation, loss function terms, and post processing, our deep adversarial network (GAN) based generative models can now generate crystal structures with higher physical feasibility and expand our previous models which can only create cubic structures.