Sebastian Pokutta

Artificial Intelligence, Optimization, and Machine Learning. Professor @TUBerlin, Vice President @ZuseInstitute, Chair @MATH+, Chair @MODAL

Research Lab

My group is interested in Artificial Intelligence, Optimization, and Machine Learning and its applications. 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. Recent examples include the estimation of biomass from satellite data, entanglement and non-locality thresholds via optimization, new constructions in extremal combinatorics via AI, as well as questions around AI and creativity. [group homepage] [more about research and projects]

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 check our openings.

Select Recent Papers

(see publications for a complete list)

  1. Sharma, U., Goel, K., Dua, A., Pokutta, S., and Woodstock, Z. (2025). A note on asynchronous Projective Splitting in Julia. Preprint. [arXiv] opt
  2. Haase, J., Klessascheck, F., Mendling, J., and Pokutta, S. (2025). Sustainability via LLM Right-sizing. Preprint. [arXiv] haiimlsustainability
  3. Zimmer, M., Spiegel, C., and Pokutta, S. (2025). Compression-aware Training of Neural Networks using Frank-Wolfe. In K. Fackeldey, A. Kannan, S. Pokutta, K. Sharma, D. Walter, A. Walther, and M. Weiser (Eds.), Mathematical Optimization for Machine Learning (pp. 137–168). De Gruyter. [PDF] [arXiv] mloptsparsity
  4. Troppens, H., Besançon, M., Wilken, S. E., and Pokutta, S. (2025). Mixed-Integer Optimization for Loopless Flux Distributions in Metabolic Networks. To Appear in Proceedings of Proceedings of Symposium on Experimental Algorithms (SEA) 2025. [arXiv] ai4sciencebiochemistryopt
  5. Haase, J., Hanel, P. H. P., and Pokutta, S. (2025). Has the Creativity of Large-Language Models peaked? An analysis of inter- and intra-LLM variability. Preprint. [arXiv] haiimlsocial
  6. Takahashi, S., Pokutta, S., and Takeda, A. (2025). Accelerated Convergence of Frank–Wolfe Algorithms with Adaptive Bregman Step-Size Strategy. Preprint. [arXiv] fwopt
  7. Carderera, A., Pokutta, S., Schütte, C., and Weiser, M. (2025). An efficient first-order conditional gradient algorithm in data-driven sparse identification of nonlinear dynamics to solve sparse recovery problems under noise. To Appear in Journal of Computational and Applied Mathematics. [PDF] [arXiv] [summary] ai4sciencemlopt
  8. Aigner, K.-M., Denzler, S., Liers, F., Pokutta, S., and Sharma, K. (2025). Scenario Reduction for Distributionally Robust Optimization. Preprint. [arXiv] optrobopt
  9. Głuch, G., Turan, B., Nagarajan, S. G., and Pokutta, S. (2025). The Good, the Bad and the Ugly: Watermarks, Transferable Attacks and Adversarial Defenses. Proceedings of ICLR 2025 Workshop on GenAI Watermarking (WMARK). [arXiv] [summary] [poster] [conference] mlxai
  10. Lasby, M., Zimmer, M., Pokutta, S., and Schultheis, E. (2025). Compressed sparse tiles for memory-efficient unstructured and semi-structured sparsity. Proceedings of ICLR 2025 Workshop on Sparsity in LLMs (SLLM). [PDF] [conference] hpcml
  11. Pauls, J., Zimmer, M., Turan, B., Saatchi, S., Ciais, P., Pokutta, S., and Gieseke, F. (2025). Capturing Temporal Dynamics in Large-Scale Canopy Tree Height Estimation. Preprint. [arXiv] [visuals] ai4sciencemlsustainability
  12. Hendrych, D., Besançon, M., Martínez-Rubio, D., and Pokutta, S. (2025). Secant Line Search for Frank-Wolfe Algorithms. Preprint. [arXiv] opt
  13. Martínez-Rubio, D., and Pokutta, S. (2025). Beyond Short Steps in Frank-Wolfe Algorithms. Preprint. [arXiv] mlopt
  14. Fackeldey, K., Kannan, A., Pokutta, S., Sharma, K., Walter, D., Walther, A., and Weiser, M. (Eds.). (2025). Mathematical Optimization for Machine Learning. de Gruyter. mlopt (Proceedings of MATH+ TES Summer Semester 2023)
  15. Pelleriti, N., Zimmer, M., Wirth, E., and Pokutta, S. (2025). Approximating Latent Manifolds in Neural Networks via Vanishing Ideals. Preprint. [arXiv] compalgmltheory
  16. Besançon, M., Designolle, S., Halbey, J., Hendrych, D., Kuzinowicz, D., Pokutta, S., Troppens, H., Viladrich Herrmannsdoerfer, D., and Wirth, E. (2025). Improved algorithms and novel applications of the FrankWolfe.jl library. Preprint. [arXiv] optsoftware
  17. Sadiku, S., Wagner, M., Nagarajan, S. G., and Pokutta, S. (2025). S-CFE: Simple Counterfactual Explanations. To Appear in Proceedings of AISTATS. [arXiv] mlxai
  18. Wirth, E., Besançon, M., and Pokutta, S. (2025). The Pivoting Framework: Frank-Wolfe Algorithms with Active Set Size Control. To Appear in Proceedings of AISTATS. [arXiv] mlopt (Oral Presentation + Conference Proceedings)
  19. Martinez-Rubio, D., Roux, C., Criscitiello, C., and Pokutta, S. (2025). Accelerated Methods for Riemannian Min-Max Problems. To Appear in Proceedings of AISTATS. [arXiv] mlopt
  20. Mexi, G., Kamp, D., Shinano, Y., Pu, S., Hoen, A., Bestuzheva, K., Hojny, C., Walter, M., Pfetsch, M. E., Pokutta, S., and Koch, T. (2025). State-of-the-art Methods for Pseudo-Boolean Solving with SCIP. Preprint. [arXiv] ipoptsoftware
  21. Roux, C., Zimmer, M., and Pokutta, S. (2025). On the Byzantine-Resilience of Distillation-Based Federated Learning. To Appear in Proceedings of ICLR. [arXiv] [summary] [code] mlopt
  22. Sadiku, S., Wagner, M., and Pokutta, S. (2025). Group-wise Sparse and Explainable Adversarial Attacks. To Appear in Proceedings of ICLR. [arXiv] [poster] ml
  23. Roux, C., Martínez-Rubio, D., and Pokutta, S. (2025). Implicit Riemannian Optimism with Applications to Min-Max Problems. Preprint. [arXiv] mlopt
  24. Mundinger, K., Zimmer, M., Kiem, A., Spiegel, C., and Pokutta, S. (2025). Neural Discovery in Mathematics: Do Machines Dream of Colored Planes? Preprint. [arXiv] ai4sciencedggraphs
  25. Haase, J., and Pokutta, S. (2024). Human-AI Co-Creativity: Exploring Synergies Across Levels of Creative Collaboration. Preprint. [arXiv] haiimlsocial
  26. Głuch, G., Turan, B., Nagarajan, S. G., and Pokutta, S. (2024). The Good, the Bad and the Ugly: Watermarks, Transferable Attacks and Adversarial Defenses. Preprint. [arXiv] [summary] [poster] mlxai
  27. Designolle, S., Vértesi, T., and Pokutta, S. (2024). Better bounds on Grothendieck constants of finite orders. Preprint. [arXiv] optphysicsquantum
  28. Abbas, A., Ambainis, A., Augustino, B., Bärtschi, A., Buhrman, H., Coffrin, C., Cortiana, G., Dunjko, V., Egger, D. J., Elmegreen, B. G., Franco, N., Fratini, F., Fuller, B., Gacon, J., Gonciulea, C., Gribling, S., Gupta, S., Hadfield, S., Heese, R., … Zoufal, C. (2024). Quantum Optimization: Potential, Challenges, and the Path Forward. Nature Reviews Physics. [PDF] [arXiv] optphysicsquantumsurvey
  29. Wirth, E., Peña, J., and Pokutta, S. (2024). Fast Convergence of Frank-Wolfe algorithms on polytopes. Preprint. [arXiv] mlopt
  30. Pauls, J., Zimmer, M., Kelly, U. M., Schwartz, M., Saatchi, S., Ciais, P., Pokutta, S., Brandt, M., and Gieseke, F. (2024). Estimating Canopy Height at Scale. Proceedings of ICML. [arXiv] [code] [visuals] ai4sciencemlsustainability
  31. Mundinger, K., Zimmer, M., and Pokutta, S. (2024). Neural Parameter Regression for Explicit Representations of PDE Solution Operators. ICLR 2024 Workshop on AI4DifferentialEquations In Science. [PDF] [arXiv] [slides] [poster] ai4scienceml
  32. Pokutta, S. (2024). The Frank-Wolfe algorithm: a short introduction. Jahresbericht Der Deutschen Mathematiker-Vereinigung, 126, 3–35. [PDF] [arXiv] mlopt
  33. Braun, G., Carderera, A., Combettes, C. W., Hassani, H., Karbasi, A., Mokthari, A., and Pokutta, S. (2022). Conditional Gradient Methods. Preprint. [arXiv] mloptsurvey

Select Recent Talks and Teaching

Recent Blog Posts

Select Outreach

News