Researchers Bettina Könighofer and Anna Galler of Graz University of Technology will each receive an Astra Prize from the Austrian Science Fund (FWF). The award comes with a grant of one million euros each over a period of five years and is intended to support long-term research projects at the international level.

Bettina Könighofer conducts research on trustworthy artificial intelligence at the Institute of Information Security. Anna Galler studies quantum materials for applications in electronics at the Institute for Theoretical Physics – Computational Physics.

Safeguards for Autonomous AI

As part of the “SEAL – Runtime Verification for Trustworthy AI” project, Bettina Könighofer is developing protective mechanisms that monitor learning AI systems during operation. These so-called “shields” are designed to prevent risky or non-compliant actions and ensure that decisions are transparent. The goal is to strengthen trust in autonomous systems such as vehicles or robots.

In her dissertation, Könighofer had already presented an approach to provably correct machine learning in which symbolic AI verifies the decisions made by neural networks. In this new project, she is expanding this approach beyond safety considerations to include ethics, fairness, and explainability. The project is part of the Austrian Bilateral AI Cluster of Excellence.

“When AI acts on its own, we must be able to trust that it will act safely and respect our rules and standards,” says Bettina Könighofer.

Quantum Materials for the Electronics of the Future

In her project “Electron Dynamics and Correlation in 2D Quantum Materials,” Anna Galler focuses on two-dimensional quantum materials consisting of just one or a few atomic layers. Due to their electronic, optical, and magnetic properties, these materials are considered a promising foundation for future electronics, sensors, and storage media.

As part of the project, Galler is developing theoretical and computer-aided quantum many-particle methods to predict the properties of these materials even before they are synthesized in the lab. She is also investigating how their behavior can be specifically influenced using light.

Since there is currently not enough data on these materials to support the use of machine learning, new computational methods running on high-performance computers are intended to close the existing research gap.

Award for Cutting-Edge Research

According to the FWF, the Astra Awards are intended to enable outstanding female and male researchers to further develop their research profiles and qualify for academic leadership positions.