Materials Generation in the Era of Artificial Intelligence: A Comprehensive Survey

TL;DR

The paper surveys AI-driven materials generation, detailing crystal representations, symmetry considerations, and the principal generative-model families (VAEs, GANs, diffusion, and autoregressive). It provides a systematic taxonomy of methods (including hybrid approaches), surveys open datasets and evaluation metrics, and discusses practical challenges like synthesizability and cross-scale modeling. By linking representations to backbones and benchmarks, the work offers a roadmap for robust, physics-informed generation of novel materials with target properties. The survey emphasizes symmetry, space-group constraints, and scalable data resources as key enablers for accelerating experimental validation and real-world deployment in materials science.

Abstract

Materials are the foundation of modern society, underpinning advancements in energy, electronics, healthcare, transportation, and infrastructure. The ability to discover and design new materials with tailored properties is critical to solving some of the most pressing global challenges. In recent years, the growing availability of high-quality materials data combined with rapid advances in Artificial Intelligence (AI) has opened new opportunities for accelerating materials discovery. Data-driven generative models provide a powerful tool for materials design by directly create novel materials that satisfy predefined property requirements. Despite the proliferation of related work, there remains a notable lack of up-to-date and systematic surveys in this area. To fill this gap, this paper provides a comprehensive overview of recent progress in AI-driven materials generation. We first organize various types of materials and illustrate multiple representations of crystalline materials. We then provide a detailed summary and taxonomy of current AI-driven materials generation approaches. Furthermore, we discuss the common evaluation metrics and summarize open-source codes and benchmark datasets. Finally, we conclude with potential future directions and challenges in this fast-growing field. The related sources can be found at https://github.com/ZhixunLEE/Awesome-AI-for-Materials-Generation.

Figures (7)

• Figure 1: A theoretically novel crystal structure should be standardized, unique, and pass both thermodynamic and kinetic stability tests. Such a structure holds the potential to revolutionize society by enabling the development of new functional materials, catalysts and energy-related materials. The generated structure provides an initial topology for diffraction-based techniques to confirm unknown crystal structures.
• Figure 2: Universal crystal structure representation using geometric graphs, textual data (Crystallographic Information File, CIF), SLICES string, and diffraction patterns (X-ray, electron, and neutron).
• Figure 3: The illustration of crystals constructed using the asu and space group.
• Figure 4: The five types of representation invariance of crystals.
• Figure 5: A schematic diagram of mainstream generative model frameworks in material generation.