The book “Artificial Intelligence and Intelligent Matter” by Michael Te Vrugt (editor) has been published by Springer in January 2026.

Here is a link.

TMM's contribution is the chapter on “Generative deep learning for the inverse design of materials” written by Yixuan Zhang, Dr. Teng Long, alumni who graduated in 2022 and now Professor at the School of Materials Science & Engineering, Shandong Universiy, Jinan, China and Prof. Hongbin Zhang.

It is a review on how generative deep learning drives the “inverse design” of materials: no longer just performing forward predictions from structure to properties, but establishing sampleable and controllable generative mechanisms on the structure-property mapping, thereby directly proposing candidate materials that meet target properties in high-dimensional spaces.

The paper consists of the three parts: first, mapping crystal structures or microstructures to a continuous latent space using appropriate representation learning, enabling the compression and reconstruction of structural distributions; subsequently, using generative methods such as VAE, GAN, and diffusion models to learn and sample structural distributions in the latent space; and finally, aligning the generative direction to the target property region through property constraints (filtering, optimization, or conditional), outputting decodable structural/microstructural candidates. These can form a closed-loop iteration with computational simulations or experiments to improve generation efficiency and design hit rates.