2026
March 2026
THK AI Journal Club established
March 2026
The THK AI Journal Club is a regular meeting where researchers and students present and discuss recent papers in AI and machine learning. The club aims to foster a collaborative environment for knowledge sharing and discussion.
February 2026
New Lectures: MLOps and Distributed Systems
February 2026
The AI for Science Group at TH Köln offers new lectures on MLOps (MLO) and Distributed Systems (BVS2). These courses cover machine learning operations for production-ready ML systems and fundamentals of distributed computing architectures, respectively.
January 2026
ICLR 2026: OrthoRF Paper Accepted
January 2026
Our paper “OrthoRF: Exploring Orthogonality in Object-Centric Representations” has been accepted at ICLR 2026! This work introduces competitive binding and orthogonality loss for synchrony-based object-centric learning, with applications in semiconductor metrology. Joint work with UvA and ASML.
New Preprint: Data-Driven Qubit Characterization
January 2026
New preprint on arXiv: “Data-Driven Qubit Characterization and Optimal Control using Deep Learning” by Paul Surrey, Julian D. Teske, Tobias Hangleiter, Hendrik Bluhm, and Pascal Cerfontaine.
2025
January 2025
AI for Science Group Founded at TH Köln
January 2025
Prof. Dr. Pascal Cerfontaine joins TH Köln as Full Professor, founding the AI for Science Group. The group focuses on physics-informed generative AI for applications in semiconductor metrology, quantum computing, biology, and ML engineering.
KIMO-QUANT Project Funded
2025
The KIMO-QUANT project (~590k€, HAW-ForschungsAkzente, 2027–2030) has been funded. The project develops physics-informed generative AI for quantum hardware optimization, focusing on electron shuttling in semiconductor spin qubits. Partners: ML4Q Cluster (RWTH, Uni Köln, FZ Juelich).