Research

Topics

Machine Learning: Clustering, Ordering, Graphical Models
Computer Vision: Image Segmentation, Multiple Object Tracking, Biomedical Image Analysis
Applied Mathematics: Combinatorial Optimization

Focus on methods and foundations: Andres B., Irmai J. and Naumann L. F. Chorded cycle facets of the clique partitioning polytope. Discrete Applied Mathematics 378:662-670, 2026
Focus on practicality and applications: Irmai J., Zhao S., Schöne M., Presberger J. and Andres B. A Graph Multi-separator Problem for Image Segmentation. Journal of Mathematical Imaging and Vision 66(5):839-872, 2024

2025-07: Jannik Irmai is showing a characterization of the chorded cycle facets of the clique partitioning polytope at EUROMIP 2025
2023-07: Bjoern Andres is talking about two generalizations of a multicut problem in the Seminar on Statistical Learning and Probability/Dynamical Systems of the MPI for Mathematics in the Sciences and the Institute of Mathematics of Leipzig University.
2023-06: Silvia Di Gregorio has accepted a position as Maître de Conférences in Combinatorial Optimization of the Université Sorbonne Paris Nord at the Laboratoire d'Informatique de Paris-Nord.
2023-05: Silvia Di Gregorio is giving an invited talk on Partial Optimality in Cubic Correlation Clustering at MIP 2023.
2022-10: Bjoern Andres is talking about lifted multicuts and image analysis. Dresdner Mathematisches Seminar.
2022-09: Jannik Irmai has won an Outstanding Master's Thesis Award by the Gesellschaft für Operations Research for work he did with Christoph Buchheim at TU Dortmund.
2022-06: Silvia Di Gregorio is presenting A Polyhedral Study of Lifted Multicuts as a poster at IPCO 2022.
2022-05: Jannik Irmai is presenting A Polyhedral Study of Lifted Multicuts as a poster at MIP 2022.
2022-02: Bjoern Andres is talking about the unreasonable effectiveness of physics in the field of computer vision in the Physics Colloquium at Heidelberg University.
2021-10: Silvia Di Gregorio is talking about joint work with Jan-Hendrik Lange on the lifted multicut problem on trees at INFORMS 2021.