Prof. Dr. Sebastian Pokutta

Vice President and Division Head
Mathematical Algorithmic Intelligence
AI in Society, Science, and Technology (AIS²T)
Zuse Institute Berlin (ZIB)

Professor for Optimization and Machine Learning
Institute of Mathematics
Electrical Engineering and Computer Science (courtesy)
Technische Universität Berlin

Research Lab. My group is interested in Artificial Intelligence, Optimization, and Machine Learning. We develop new methodologies (e.g., new optimization and learning algorithms), work on combining learning and decision-making, as well as design AI Systems for real-world deployment in various application contexts. [more]

(Informal) TL;DR. We use computers to learn from data and make better decisions.

Prospective Students. If you are interested in working in our group or writing your MS/BS thesis please only use the email applications-aisst@zib.de.

Recent Papers.

1. Kossen, T., Hirzel, M. A., Madai, V. I., Boenisch, F., Hennemuth, A., Hildebrand, K., Pokutta, S., Sharma, K., Hilbert, A., Sobesky, J., Galinovic, I., Khalil, A. A., Fiebach, J. B., and Frey, D. (2021). Towards sharing brain images: Differentially private TOF-MRA images with segmentation labels using generative adversarial networks. Preprint.
2. Zimmer, M., Spiegel, C., and Pokutta, S. (2021). Back to Basics: Efficient Network Compression via IMP. Preprint. [arXiv] [code]
3. Tsuji, K., Tanaka, K., and Pokutta, S. (2021). Sparser Kernel Herding with Pairwise Conditional Gradients without Swap Steps. Preprint. [arXiv]
4. Macdonald, J., Besançon, M., and Pokutta, S. (2021). Interpretable Neural Networks with Frank-Wolfe: Sparse Relevance Maps and Relevance Orderings. Preprint. [arXiv]
5. Criado, F., Martinez-Rubio, D., and Pokutta, S. (2021). Fast Algorithms for Packing Proportional Fairness and its Dual. Preprint. [arXiv] [poster]
6. Sofranac, B., Gleixner, A., and Pokutta, S. (2021). An Algorithm-Independent Measure of Progress for Linear Constraint Propagation. To Appear in Proceedings of International Conference on Principles and Practice of Constraint Programming. [arXiv] [video]
7. Carderera, A., Besançon, M., and Pokutta, S. (2021). Simple steps are all you need: Frank-Wolfe and generalized self-concordant functions. To Appear in Proceedings of NeurIPS. [arXiv] [slides] [poster] [code]
8. Roux, C., Wirth, E., Pokutta, S., and Kerdreux, T. (2021). Efficient Online-Bandit Strategies for Minimax Learning Problems. Preprint. [arXiv]
9. Besançon, M., Carderera, A., and Pokutta, S. (2021). FrankWolfe.jl: a high-performance and flexible toolbox for Frank-Wolfe algorithms and Conditional Gradients. Preprint. [arXiv] [summary] [slides] [code]
10. Chmiela, A., Khalil, E., Gleixner, A., Lodi, A., and Pokutta, S. (2021). Learning to Schedule Heuristics in Branch-and-Bound. To Appear in Proceedings of NeurIPS. [arXiv] [summary] [poster]
11. Kerdreux, T., Roux, C., d’Aspremont, A., and Pokutta, S. (2021). Linear Bandits on Uniformly Convex Sets. To Appear in Journal of Machine Learning Research (JMLR). [arXiv] [summary]
12. Kerdreux, T., d’Aspremont, A., and Pokutta, S. (2021). Local and Global Uniform Convexity Conditions. Preprint. [arXiv]
13. Carderera, A., Diakonikolas, J., Lin, C. Y., and Pokutta, S. (2021). Parameter-free Locally Accelerated Conditional Gradients. To Appear in Proceedings of ICML. [arXiv] [slides]
14. Pokutta, S. (2021). Mathematik, Machine Learning und Artificial Intelligence. To Appear in Mitteilungen Der DMV (German). [PDF]
15. Braun, G., and Pokutta, S. (2021). Dual Prices for Frank-Wolfe Algorithms. Preprint. [arXiv]
16. Carderera, A., Pokutta, S., Schütte, C., and Weiser, M. (2021). CINDy: Conditional gradient-based Identification of Non-linear Dynamics – Noise-robust recovery. Preprint. [arXiv] [summary]
17. Kerdreux, T., d’Aspremont, A., and Pokutta, S. (2021). Projection-Free Optimization on Uniformly Convex Sets. To Appear in Proceedings of AISTATS. [arXiv] [summary] [slides]
18. Combettes, C. W., and Pokutta, S. (2021). Complexity of Linear Minimization and Projection on Some Sets. To Appear in Operations Research Letters. [arXiv] [code]
19. Combettes, C. W., Spiegel, C., and Pokutta, S. (2020). Projection-Free Adaptive Gradients for Large-Scale Optimization. Preprint. [arXiv] [summary] [code]

Select Recent Talks and Teaching.

• 11/2021: (technical) “Discrete Optimization in Machine Learning - an (informal) overview”. Talk at Oberwolfach Workshop on Combinatorial Optimization (Oberwolfach). [slides]
• 10/2021: (technical) “Fast algorithms for 1-fair packing (and its dual)”. Talk at HIM Workshop: Continuous approaches to discrete optimization (Bonn). [slides]
• 09/2021: (technical) “Conditional Gradients - a tour d’horizon”. Talk at AI Campus Berlin Tech Lunch Talk (online). [slides]
• 02/2021: (technical) “Structured ML Training via Conditional Gradients”. Talk at IPAM Deep Learning and Combinatorial Optimization Workshop (online). [slides] [video]
• 11/2020: (technical) “Conditional Gradients: Overview and Recent Advances”. Talk at RWTH Aachen Mathematical Colloquium (online). [slides]
• SoSe/2021: Discrete Optimization and Machine Learning (seminar)

Recent Blog Posts.

• 10/2021: Simple steps are all you need
• 06/2021: New(!!) NeurIPS 2021 competition: Machine Learning for Discrete Optimization (ML4CO)
• 05/2021: Learning to Schedule Heuristics in Branch and Bound
• 04/2021: FrankWolfe.jl: A high-performance and flexible toolbox for Conditional Gradients
• 04/2021: Linear Bandits on Uniformly Convex Sets

News.

• 10/2021: Math+ Cluster presentation at the Humboldt Forum “Mit Mathematik die Welt verbessern?” (German) [video]
• 10/2021: Our group received a Google Research Award to explore the learning of heuristic schedules in MIP solvers.
• 10/2021: One project funded by the Math+ Research Center.
• 11/2020: Our group received a Google Research Award to support our work on Integer Programming solvers.
• 10/2020: Four projects funded by the Math+ Research Center.