Speaker
Description
Sequential active learning (SAL)-driven workflows can efficiently guide experiments and simulations towards the discovery of materials with desired properties [1]. However, AI and machine-learning approaches commonly used in these workflows rely on the knowledge of key physical parameters describing the materials property of interest. These low-dimensional representations are typically unknown. Here, we address this challenge by developing a SAL workflow based on the sure-independence screening and sparsifying operator (SISSO) approach [2,3]. SISSO identifies, even based on moderate amounts of data, models for materials properties as analytical expressions depending on key parameters, out of many offered ones. Crucially, we train ensembles of SISSO models in order to obtain not only mean predictions but also to quantify the uncertainty of the predictions, which are used to navigate the previously unexplored regions of materials space [4]. We demonstrate the SISSO-guided workflow by identifying acid-stable oxides for the water-splitting reaction by using high-quality DFT-HSE06 calculations [5].
[1] J. H. Montoya, K. T. Winther, R. A. Flores, T. Bligaard, J. S. Hummelshøj, M. Aykol, Autonomous intelligent agents for accelerated materials discovery, Chem. Sci. 11, 2022; https://doi.org/10.1039/D0SC01101K
[2] R. Ouyang, S. Curtarolo, E. Ahmetcik, M. Scheffler, 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; https://doi.org/10.1103/PhysRevMaterials.2.083802
[3] T. A. R. Purcell, M. Scheffler, C. Carbogno, L. M. Ghiringhelli, SISSO++: A C++ Implementation of the Sure-Independence Screening and Sparsifying Operator Approach, J. Open Source Softw. 7, 3960, 2022; https://doi.org/10.21105/joss.03960
[4] J. Behler, G. Csanyi, L. Foppa , K. Kang, M. F. Langer, J. T. Margraf, A. S. Nair, T. A. R. Purcell, P. Rinke, M. Scheffler, A. Tkatchenko, M.Todorovic, O. T. Unke, Y. Yao, Workflows for Artificial Intelligence, ChemRxiv, 2024; https://doi.org/10.26434/chemrxiv-2024-vw06p
[5] A. S. Nair, AI-guided Workflow for the Discovery of Acid-Stable Oxides, https://gitlab.com/akhilsnair/sl-sisso
