AI guided workflows for efficiently screening the materials space

Matthias Scheffler

doi:10.61109/cs.202403.129

Artificial intelligence (AI) may capture the properties and functions of materials better than previous theoretical/computational methods because it targets correlations and does not assume a single, specific underlying physical model. Therefore, it addresses the full intricacy of the numerous processes that govern the function of materials. However, the statistical analysis and interpretation of AI models require careful attention.
The review article started with a brief discussion of historical aspects of data-centric science. It then focused on the recently developed, explainable AI methods ^[8,10] and applications ^[2,11,12]. The identified "rules" determine the properties and functions of materials. The rules depend on descriptive parameters called "materials genes." As genes in biology, they are correlated with a certain material property or function. Thus, these materials genes help to identify materials that are, for example, better electrical conductors or better heat insulators or better catalysts.

Keywords:

artificial intelligence

machine learning

active learning

symbolic regression

materials science

materials genes

This project was done together with T.A.R. Purcell, L. Foppa, M. Boley and several others. It was supported by the ERC Advanced Grant TEC1p (European Research Council, Grant Agreement No. 740233).

1. M. Scheffler, M. Aeschlimann, M. Albrecht, T. Bereau, H.J. Bungartz, C. Felser, M. Greiner, A. Groß, C.T. Koch, K. Kremer, W.E. Nagel, M. Scheidgen, C. Wöll, and C. Draxl, FAIR data enabling new horizons for materials research, Nature 604, 635 (2022).

2. T.A.R. Purcell, M. Scheffler, L.M. Ghiringhelli, and C. Carbogno, Accelerating materials-space exploration for thermal insulators by mapping materials properties via artificial intelligence, npj Comput. Mater. 9, 112 (2023).

3. F. Knoop, T.A.R. Purcell, M. Scheffler, and C. Carbogno, Anharmonicity in thermal insulators: an analysis from first principles, Phys. Rev. Lett. 130, 236301 (2023).

4. F. Knoop, M. Scheffler, and C. Carbogno, Ab initio Green-Kubo simulations of heat transport in solids: method and implementation, Phys. Rev. B 107, 224304 (2023).

5. D.L. Perry, Handbook of Inorganic Compounds (CRC Press, 2016).

6. A. Togo, L. Chaput, and I. Tanaka, Distributions of phonon lifetimes in Brillouin zones, Phys. Rev. B 91, 094306 (2015).

7. F. Knoop, T.A.R. Purcell, M. Scheffler, and C. Carbogno, Anharmonicity measure for materials, Phys. Rev. Mater. 4, 083809 (2020).

8. R. Ouyang, S. Curtarolo, E. Ahmetcik, M. Scheffler, and L.M. Ghiringhelli, SISSO: a compressed-sensing method for identifying the best low-dimensional descriptor in an immensity of offered candidates, Phys. Rev. Mater. 2, 083802 (2018).

9. M. Boley, F. Luong, S. Teshuva, D.F. Schmidt, L. Foppa, and M. Scheffler, From prediction to action: critical role of performance estimation for machine-learning-driven materials discovery, arXiv:2311.15549 (2023).

10. B.R. Goldsmith, M. Boley, J. Vreeken, M. Scheffler, and L.M. Ghiringhelli, Uncovering structure-property relationships of materials by subgroup discovery. New J. Phys. 19, 013031 (2017).

11. L. Foppa, L.M. Ghiringhelli, F. Girgsdies, M. Hashagen, P. Kube, M. Hävecker, S.J. Carey, A. Tarasov, P. Kraus, F. Rosowski, R. Schlögl, A. Trunschke, and M. Scheffler, Materials genes of heterogeneous catalysis from clean experiments and artificial intelligence. MRS Bull. 46, 1016 (2021).

12. L. Foppa, C. Sutton, L.M. Ghiringhelli, S. De, P. Löser, S.A. Schunk, A. Schäfer, and M. Scheffler, Learning design rules for selective oxidation catalysts from high-throughput experimentation and artificial intelligence, ACS Catal. 12, 2223 (2022).