|
Foundations of Artificial Intelligence (FAI) Group
|
|
I did my undergraduate studies at the Australian National University, Canberra, and have been in Saarbrücken since 2022. I was part of IMPRS TRUST for two years supervised by Prof. Christoph Weidenbach.
Now, I am a PhD student at the chair of Prof. Jörg Hoffmann in the Computer Science department at Saarland University. I am also a member of the Neuroexplicit Models of Language, Vision and Action research training group. A short CV can be found here (PDF).
My primary research interests lie in the representation, learning, verification, and repair of action policies in nondeterministic planning environments. Specifically, I am focused on understanding possible policy representations — ranging from symbolic formulations to machine learning models such as neural networks, tree ensembles, or hybrid neuro-symbolic architectures. A key challenge is to balance expressiveness and verifiability across these representations. Once policies are synthesized or learned, I aim to formally verify their behavior with respect to safety properties. If a policy is found to be unsafe, I want to analyze and localize the source of the unsafe behavior and develop techniques for policy repair or retraining with guarantees of improved safety.