Tobias Ladner

I am a PhD student at the Technical University of Munich, specializing in AI safety with a focus on the formal verification of neural networks and cyber-physical systems. My research is dedicated to ensuring that AI systems perform reliably in safety-critical applications.

A list of publications is available below. For collaboration inquiries or more information, please email me at tobias.ladner@tum.de or visit my university profile.

I am also the administrator of CORA — a toolbox for continuous reachability analysis — which supports my work in formal neural network verification (and more!). CORA regularly competes in competitions involving the formal verification of neural networks (VNN-COMP, ARCH-COMP), in which I both serve as evaluation and repeatability chair. As part of my duties at the Technical University of Munich, I supervise student theses, design seminars and practical courses, as well as give lectures to students about my research. My work is supported by the project "Formal Verification of Analog AI Hardware (FAI)" funded by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG).

• [September 2025] I am excited to announce my 6-month research stay at Prof. Yasser Shoukry's lab at UC Irvine, USA.
• [May 2025] Our paper, "Explaining, Fast and Slow: Abstraction and Refinement of Provable Explanations", was accepted by the International Conference on Machine Learning (ICML).
• [April 2025] Our paper, "Formal Verification of Graph Convolutional Networks with Uncertain Node Features and Uncertain Graph Structure", was accepted by the Transactions on Machine Learning Research (TMLR).