Foundation Model for Material Science


  • Seiji Takeda IBM Research - Tokyo
  • Akihiro Kishimoto IBM Research - Tokyo
  • Lisa Hamada IBM Research - Tokyo
  • Daiju Nakano IBM Research - Tokyo
  • John R. Smith IBM Thomas J. Watson Research Center



Machine Learning, Foundation Model, Cheminformatics, Material Discovery


Foundation models (FMs) are achieving remarkable successes to realize complex downstream tasks in domains including natural language and visions. In this paper, we propose building an FM for material science, which is trained with massive data across a wide variety of material domains and data modalities. Nowadays machine learning models play key roles in material discovery, particularly for property prediction and structure generation. However, those models have been independently developed to address only specific tasks without sharing more global knowledge. Development of an FM for material science will enable overarching modeling across material domains and data modalities by sharing their feature representations. We discuss fundamental challenges and required technologies to build an FM from the aspects of data preparation, model development, and downstream tasks.




How to Cite

Takeda, S., Kishimoto, A., Hamada, L., Nakano, D., & Smith, J. R. (2023). Foundation Model for Material Science. Proceedings of the AAAI Conference on Artificial Intelligence, 37(13), 15376-15383.