Dresden 2026 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

FM: Fachverband Funktionsmaterialien

FM 6: Focus Session: Materials Discovery I – Material informatics

FM 6.1: Hauptvortrag

Dienstag, 10. März 2026, 09:30–10:00, BEY/0138

Robust Data Generation, Heuristics and Machine Learning for Materials Design — •Janine George — BAM Berlin, Germany — University of Jena

Machine learning (ML) offers new routes to overcome the limitations of density functional theory (DFT) for advanced materials. We present data-generation strategies and workflows for ML interatomic potentials, including large-scale quantum-chemical bonding analysis.[1,2,3] Incorporating bonding descriptors into ML models enables prediction of phononic properties and validation of correlations between bonding strength, force constants, and thermal conductivity.[3] We introduce autoplex, an automated framework for training ML potentials, supporting general-purpose and phonon-focused workflows.[4] These developments provide a basis for fine-tuning foundation models for thermal transport at reduced cost.[5] For properties such as magnetism or synthesizability, we discuss complementary approaches, comparing ab initio methods with chemical heuristics and experimental data-driven ML models.[6,7] Our work advances scalable, accurate simulations for materials discovery.

References: [1] M. K. Horton, et al. Nat. Mater. 2025, 24, 1522*1532. [2] A. M. Ganose, et al., Digit. Discov. 2025, 4, 1944*1973. [3] A. A. Naik, et al. Sci. Data 2023, 10, 610. [4] Y. Liu, et al. Nat. Commun. 2025, 16, 7666. [5] J. Bustamante, et al. 2025, DOI 10.48550/arXiv.2510.23133. [6] S. Amariamir, et al. Digit. Discov. 2025, 4, 1437*1448. [7] K. Ueltzen, et al. 2025, DOI 10.26434/chemrxiv-2025-xj84d.

Keywords: high throughput; machine learning; machine-learned interatomic potentials; heuristics