Regensburg 2025 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz
AKPIK 2: Machine Learning Prediction and Optimization Tasks
AKPIK 2.3: Vortrag
Dienstag, 18. März 2025, 10:00–10:15, H5
Co-orchestration of multiple instruments for automated exploration of structure-property relationships in combinatorial libraries — •Boris Slautin1, Utkarsh Pratiush2, Ilia Ivanov3, Yongtao Liu3, Rohit Pant4, Xiaohang Zhang4, Ichiro Takeuchi4, Maxim Ziatdinov5, and Sergei Kalinin2,5 — 1University of Duisburg-Essen, Essen, Germany — 2University of Tennessee, Knoxville USA — 3Oak Ridge National Laboratory, Oak Ridge, USA — 4University of Maryland, College Park, USA — 5Pacific Northwest National Laboratory, Richland, USA
The advancement of combinatorial synthesis techniques has significantly accelerated the development of novel materials. However, closing the loop in materials design requires powerful approaches for characterizing the synthesized libraries. This is a non-trivial task, as characterization often involves revealing multiple methods.
We present a multimodal co-orchestration framework for autonomous combinatorial library characterization through the simultaneous coordination of various tools (e.g., Raman spectroscopy, SPM). The multimodal co-orchestration enables the real-time utilization of acquired knowledge about one property to accelerate the exploration of other properties measured by different methods. The capabilities of the proposed framework were validated by the co-orchestrating of the SPM and Raman techniques in the Sm-BiFeO3 combinatorial library. The workflow confirms its effectiveness in optimizing the exploration trajectory. The proposed framework is general and can be extended to multiple modalities and arbitrary dimensionality of signals.
Keywords: combinatorial libraries; Bayesian optimization; co-orchestration; automated experiment