Priv.-Doz. Dr. Bogdan Savchynskyy Group Leader Dr. Bogdan Savchynskyy HCI am IWR Berlinerstr. 43 69120 Heidelberg Room B108, Phone: +49 (6221) 54 14604 Email: bogdan.savchynskyy at iwr.uni-heidelberg.de

About me

Publications

Projects

Teaching & Presentations

Practical, Bachelor and Master Projects

I am constantly looking for motivated students for Practical, Bachelor and Master projects in the field of large-scale combinatorial optimization with applications in computer vision and machine learning. Apart from the usual formal requirements to start a project I expect

• An excellent mark for my lecture "Optimization for Machine Learning" or similar background in integer linear programming and combinatorial optimization.
• Strong programming skills in Python and C++ and ability to work with a third-party code.

The best candidates may get a HiWi position.

PhD Student Positions

I am constantly looking for motivated PhD students. Requirements are similar to those for Master students:

• An excellent mark for my lecture "Optimization for Machine Learning" or similar background in integer linear programming and combinatorial optimization.
• Strong programming skills in Python and C++ and ability to work with a third-party code.

For a successful PhD under my guidance you are required to have strong mathematical as well as programming skills. From my side I offer a tight supervision and a full-time TV-L 13 position (for unknown candidates after a trial period).

News!

2024, May Our DFG project proposal "Optimization Techniques for Multi-Graph Matching" has been accepted!
2024, Jan. Our paper about unsupervised learning of graph matching has been accepted to AAAI 2024! Check it at the Publication page!
2022, Jul. Our graph matching benchmark paper has been accepted to ECCV 2022! Check it at the Publication page as well as the benchmark web-page!
2022, Mar. Our joint together with Dagmar Kainmüller DFG project proposal "Unsupervised Model Discovery for Stereotypical Organisms" has been accepted!
2021, Jul. Our paper describing the new state-of-the-art graph matching solver is accepted to ICCV as oral! See the Publication page.
2021, Jan. Our paper describing the new state-of-the-art graph matching solver is now on ArXiv! See the Publication page.
2020, Jan.: Two our papers have been accepted to AISTATS 2020! See the Publication page.
2019, Dec.: My book Discrete Graphical Models - An Optimization Perspective is published!
2019, Sep.: We start a regular research seminar on convex and combinatorial optimization!
2018, Aug.: My book Discrete Graphical Models - An Optimization Perspective is submitted for publication to Now Publishers!
2018, July: Our paper about a new higly parallelizable state-of-the-art message passing method for dense non-Gaussian graphical models accepted to ECCV 2018! See the reference to the source code and supplementary material at the Publication page.
2017, Nov.: Our paper about a new scalable ILP method (CombiLP) accepted to AAAI 2018! See the reference to the source code and supplementary material at the Publication page.
2017, Sep.: Our overview of CNN+CRF methods was accepted to IEEE Signal Processing Magazine!