Related Experiment Videos
AI for Accelerated Materials Discovery: From Generative Design to Autonomous Realization
Jaehwan Choi1, Seongmin Kim1, Junkil Park1,2
1Department of Chemical and Biological Engineering (BK21 four), and Institute of Chemical Processes, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul08826, Korea.
Chemical Reviews
|August 12, 2026
Summary
Artificial intelligence (AI) is revolutionizing materials discovery through data-driven workflows. This review explores AI methods for designing, synthesizing, and experimentally validating new materials, accelerating scientific advancement.
Area of Science:
- Materials Science
- Computer Science
- Chemistry
Background:
- Traditional materials discovery relies on time-consuming trial-and-error methods.
- Artificial intelligence (AI) offers a paradigm shift towards automated, data-driven approaches.
- Accelerating the discovery of novel materials is crucial for technological innovation.
Purpose of the Study:
- To review emerging AI methodologies for accelerated materials discovery.
- To examine the integration of computational design, data infrastructure, synthesis planning, and autonomous experimentation.
- To highlight AI's capabilities and limitations in creating experimentally grounded materials discovery workflows.
Main Methods:
- Survey of generative models for inverse materials design (VAEs, GANs, diffusion models, etc.).
- Discussion of physics-informed and data-efficient AI strategies for low-data regimes.
- Review of multimodal foundation models integrating diverse data types (structures, text, spectra).
- Examination of AI-driven synthesis planning and autonomous laboratories.
- Addressing critical infrastructure challenges like database limitations and FAIR data principles.
Main Results:
- Generative AI models are evolving for de novo design of crystalline materials with specific properties.
- Physics-informed and multimodal AI enhance model generalizability and cross-task generalization.
- AI-driven synthesis planning and autonomous labs bridge the computational-experimental gap.
- Addressing data infrastructure challenges is key for reliable AI application in materials discovery.
Conclusions:
- AI is fundamentally transforming materials discovery towards efficient, automated workflows.
- Integrating computational design, synthesis, and experimentation via AI is essential.
- Continued development in AI methodologies and data infrastructure will further accelerate materials innovation.