GenMatLab: A Generative Platform for Inverse Materials Design

In addition to the forward inference of materials properties using machine learning, generative deep learning techniques applied on materials science allow the inverse design of materials, i.e., assessing the composition -processing -(micro-)structure -property relationships in a reversed way. In this review, we focus on the (micro-)structure -property mapping, i.e., crystal structure -intrinsic property and microstructure -extrinsic property, and summarize comprehensively how generative deep learning can be performed. Three key elements, i.e., the construction of latent spaces for both the crystal structures and microstructures, generative learning approaches, and property constraints, are discussed in detail. A perspective is given outlining the challenges of the existing methods in terms of computational resource consumption, data compatibility, and yield of generation. related to flue gas separation CDVAE[64,65] Vectors VAE Stability high multicomponent high Cond-DFC-VAE[66] Voxels VAE Formation energy, bandgap, bulk/shear modulus, etc. middle multicomponent high FTCP[67,68] Vectors VAE Formation energy, bandgap, Thermoelectric power factor high multicomponent low ZeoGAN[69] Voxels GAN Heat absorption middle ternary high Composition Conditioned Crystal GAN[70] Vectors GAN Pourbaix stability and the band gaps high ternary low PGCGM[71,72] Vectors GAN Formation energy high multicomponent high