From automata theory to trustworthy AI

What can theoretical computer science tell us about the reliability of artificial intelligence? For Dr. Chris Köcher, this question connects two fields that have shaped his research: automata and formal language theory on the one hand, and the analysis of AI models on the other.

Chris has recently joined the Verification and Formal Guarantees of Machine Learning group headed by Prof. Daniel Neider as a Postdoc at the Research Center Trustworthy Data Science and Security (RC Trust).

His research interests lie mainly in automata, formal languages, and verification. In 2022, he obtained his PhD in theoretical computer science at Technische Universität Ilmenau. In his dissertation, he studied algebraic properties of automata with different storage mechanisms, with the aim of addressing typical verification questions.

After his PhD, Chris joined the Max Planck Institute for Software Systems in Kaiserslautern for his first postdoctoral position. There, he worked on separability problems for several classes of formal languages. During this time, he also began to study large language models from the perspective of expressivity and verification – a line of research he plans to continue at RC Trust.

This makes his work a strong fit for Daniel Neider’s group, where formal methods and machine learning meet. At RC Trust, Chris will contribute to research that asks how AI systems can be better understood, analysed, and checked with mathematical rigour.

Beyond his academic work, Chris enjoys travelling, photography, and playing video and board games.

Welcome to RC Trust, Chris!