AI Foundation models and Active Learning for Materials Discovery and Process Design

Back

Xiaonan Wang1*

1Department of Chemical Engineering, Tsinghua University, Beijing, China

*Corresponding Author: wangxiaonan@tsinghua.edu.cn

As the world faces intensifying climate impacts and the urgent need to mitigate greenhouse gas emissions, the role of materials research in driving sustainable development becomes increasingly critical. Artificial intelligence (AI) has become pivotal in materials innovation and development, promising to catalyze breakthroughs by integrating with traditional materials technologies. This talk will introduce our smart systems engineering approaches, combining multi-scale modeling and learning in materials technologies and process engineering for sustainability. Our research highlights the transformative potential of integrating theoretical calculations, deep learning models, and active learning strategies for designing high-performance catalysts and functional materials, for Carbon Capture, Utilization, and Storage (CCUS) as well as clean energy applications, aiming toward a sustainable future and aligning with net-zero goals. We also developed various foundation models and tools to enhance both computational and experimental approaches to catalyst design, marking a significant step toward pre-trained models for catalyst discovery and process optimization. The talk will conclude by presenting the latest research progress on large-scale AI foundation models in the field, along with an analysis of their future potential.

Keywords: Machine Learning, AI for Science, Catalyst Design, CCUS, Smart Systems Engineering

References

1. Su, J.; Li, J.; Guo, N.; Zhang, C.; Wang, X.*; Lu, J.* Intelligent Synthesis of Magnetic Nanographenes via Chemist-Intuited Atomic Robotic Probe. Nat. Synth. 2024, 3 (4), 466–476.

2. Ge, X.; Yin, J.; Liu, X.*; Wang, X.*; Duan, X.* Atomic Design of Alkyne Semihydrogenation Catalysts via Active Learning. J. Am. Chem. Soc. 2024, 146 (7), 4993–5004.

3. Chen, H.; Zheng, Y.; Li, J.; Li, L.; Wang, X.* AI for Nanomaterials Development in Clean Energy and Carbon Capture, Utilization and Storage (CCUS). ACS Nano 2023, 17 (11), 9763–9792.

4. Yang, H.; Li, J.; Wang, X.*; Chen, P.-Y.* Automatic Strain Sensor Design via Active Learning and Data Augmentation for Soft Machines. Nat. Mach. Intell. 2022, 4 (1), 84–94.

5. Xu, S.; Li, J.; Cai, P.; Liu, X.; Liu, B.*; Wang, X.* Self-Improving Photosensitizer Discovery System via Bayesian Search with First-Principle Simulations. J. Am. Chem. Soc. 2021, 143 (47), 19769–19777.