Growing the Network—Generative AI

This chapter is devoted to an exploration of modern Generative Deep Learning models for growing the network, that are being increasingly used in the search for new materials with specific properties, and some of the most exciting developments in this area. In the process, we will bring out the threads of connections to diverse topics encountered in earlier chapters, such as similarity kernels, Gaussian processes, spectral graph theory and random matrix theory. We will see how deep learning and generative methods are being used for the design of novel materials, and how similarity kernels and chemical space representations are key to these advances.