Two researchers at Graz University of Technology will each receive €1 million over five years from the Austrian Science Fund's ASTRA awards to advance trustworthy artificial intelligence and next-generation quantum materials. Bettina Könighofer and Anna Galler lead projects that address fundamental challenges: ensuring autonomous AI systems act safely and predicting the properties of ultra-thin materials before they are synthesized.
FWF Astra Awards Support Quantum Materials and Trustworthy AI
The Austrian Science Fund designed the Astra program to back exceptional researchers expected to make substantial international contributions. Könighofer's project, SEAL - Shielding for Explainable Correctness of Learned Systems, builds on her doctoral work that produced the first approach to provably correct machine learning. Galler's project, Electron Dynamics and Correlation in 2D Quantum Materials, aims to unlock the potential of materials just one or a few atoms thick for future electronics and sensors.
Symbolic AI Creates Safety Shields for Autonomous Systems
Rather than trying to interpret the full complexity of neural networks, Könighofer's team uses symbolic AI to inspect individual decision points. "With our approach, we do not aim to understand the decision making of neural networks - that is, subsymbolic AI - with their millions of parameters as a whole, because that is an extraordinarily difficult challenge," she said. The system proactively checks for risky or non-compliant actions before execution, preventing outcomes like an autonomous vehicle driving off an embankment. "When AI acts autonomously, we must be able to trust that it will act safely and respect our rules and standards," Könighofer said. The project now extends beyond collision avoidance to include ethics, fairness, and explainability.
Computational Methods Predict 2D Quantum Material Behavior
Anna Galler is developing theoretical and numerical quantum many-particle methods to forecast electronic, optical, and magnetic properties of two-dimensional materials. Accurate prediction has remained elusive because of complex quantum mechanical interactions between electrons. "In this project, I am developing theoretical and numerical quantum many-particle methods to predict the electronic, optical and magnetic properties of these quantum materials and to identify promising new materials before they are synthesised in the laboratory," Galler said. Her work focuses on how stacking different layers can create entirely new properties, such as turning a metal into an insulator. This predictive capability could accelerate the development of ultra-thin, flexible transistors and advanced data storage.
Why this matters for science and research professionals
The ASTRA awards signal strong institutional support for high-risk, high-reward research that bridges fundamental science and practical application. Könighofer's method of shielding AI systems without full network transparency offers a pragmatic path to certifiable autonomous systems. Galler's quantum many-particle methods address a bottleneck in materials discovery, where limited data currently rules out machine learning approaches. Both projects demonstrate how targeted funding can tackle specific technical barriers that hold back entire fields.
Your membership also unlocks:
