Arthur Hilbert

Arthur Hilbert is a Researcher in the VeraXtract project at the Quality and Usability Lab (Q&U Lab) within the XplaiNLP group at the Technical University of Berlin, joining in February 2025.

Education:

• B.Eng. in Media Technology, Georg Simon Ohm University of Applied Sciences, Nuremberg (2018)
• M.Sc. in Cognitive Systems, University of Potsdam (2024); Thesis: "SchreiberLing: A Portable Model for Classification and Repair of German Speech-To-Text Data" supervised by Prof. Dr. Manfred Stede

Research Interests: Hilbert focuses on explainable AI (xAI) with core work in disinformation narrative emergence, user-centric xAI development, mechanistic interpretability, and inductive logic programming applications. His secondary interests span AI-driven sports analytics, game-based AI systems, higher-dimensional generative adversarial networks, and neural cellular automata architectures, reflecting interdisciplinary exploration in AI safety and novel applications.

Professional Background: Prior to academia, he developed professional audio plugins at MAGIX Software. During master's studies, he contributed to key entity recognition in resource-poor domains using LLMs at Intel Labs' Multimodal Dialogue group and advanced the open-source 3D analysis tool webknossos.org at Scalable Minds.

Advising and Projects: As a current PhD researcher, he does not supervise students. His VeraXtract project work centers on foundational AI interpretability research within the Q&U Lab's framework, though specific grants remain unmentioned in source materials.

Research Environment: He operates within the XplaiNLP group focusing on human-AI interaction quality, leveraging both industry experience and academic training to bridge theoretical AI development with practical usability concerns.