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

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 4: Focus: Applications of Deep Neural Networks

AKPIK 4.2: Hauptvortrag

Dienstag, 18. März 2025, 14:30–15:00, H5

Inverse Design in Electromagnetics with Artificial Intelligence — •Willie Padilla — Duke University, Durham, North Carolina, USA

Artificial electromagnetic materials (AEMs) have enabled exotic electromagnetic responses that are difficult or impossible to achieve with naturally occurring materials. However, as AEMs have become more complex, the relationship between their structure and resulting properties is increasingly less understood, or sometimes completely unknown. Deep neural networks (DNNs) have been shown to effectively infer the relationship between AEM geometry and their electromagnetic properties, using simulated training data. More recently, a type of DNN * termed a large language model (LLM) * has shown a remarkable ability to respond to complex prompts. This presentation explores the potential of DNNs and LLMs for the inverse design of AEMs. I present a LLM fine-tuned on simulated data that can predict electromagnetic spectra over a range of frequencies given a text prompt that only specifies the AEM geometry. In view of the great potential of deep learning for the future of AEM research, we review the status of the field, focusing on recent advances, open challenges, and future directions.

Keywords: Electromagnetics; Inverse Design; Metamaterials; Deep Learning; Artificial Intelligence