Stefan Heidekrüger, M.Sc.

E-Mail: stefan dot heidekrueger at in.tum.de 

Phone: +49 (0) 89 289 - 17530
Fax: +49 (0) 89 289 - 17535
Office: Room 01.10.056
Boltzmannstr. 3
85748 Munich, Germany

Hours: by arrangement


Short Bio

I'm a PhD student at the DSS chair supervised by Prof. Bichler. My research focusses on computation of equilibria in incomplete information games, especially markets and auctions and using multi-agent reinforcement learning methods. You can find more information about me on GitHub or LinkedIn.

 

Since 2018  Doctoral Student and Research Associate, Decision Sciences & Systems, TUM
2016 - 2018  Data Scientist, Business Analytics and Artificial Intelligence, Telefónica Germany
2014 - 2016 M.Sc. Mathematics in Operations Research, Technische Universität München
2014 Erasmus+ student at KTH Royal Institute of Technology (Stockholm, Sweden)
2013 - 2016 internships at a.hartrodt (2013) and zeb Consulting (2015)
working student positions at a.hartrodt (2013-14), Telefónica Germany (2016), and SAP (2016)
student research assistant positions at TUM (2014, 2015) and HelmholtzZentrum München (2015-16)
2012 - 2013 Visiting Student at The Hong Kong University of Science and Technology 
2010 - 2014 B.Sc. Mathematics, TUM

Publications

Journal Papers

Learning equilibria in symmetric auction games using artificial neural networks.
M. Bichler, M. Fichtl, S. Heidekrüger, N. Kohring, and P. Sutterer.  Nature Machine Intelligence,  3, 687–695 (2021). [link | unedited authors' manuscript | supplement ]
A previous version was presented at the 2020 annual meeting of the NBER Working Group on Market Design: [pdf ]

Conference Proceedings

Approaching the Overbidding Puzzle in All-Pay Auctions: Explaining Human Behavior through Bayesian Optimization and Equilibrium Learning
M. Ewert, S. Heidekrüger, and M. Bichler. Extended abstract to appear in the Proceedings of the 21st International Conference on Autonomous Agents and Multiagent Learning, May 2022. (AAMAS-22) (forthcoming).

Equilibrium Learning in Auction Markets.
S. Heidekrüger, Proceedings of the Twenty-Seventh AAAI/SIGAI Doctoral Consortium (AAAI-DC-22), February 2022 (forthcoming)

Peer-Reviewed Workshop Papers

Equilibrium Computation for Auction Games via Multi-Swarm Optimization
N. Kohring, C. Froehlich, S. Heidekrüger, and M. Bichler
Accepted to the 2022 AAAI Workshop on Reinforcement Learning in Games (AAAI-RLG-22)

Equilibrium Learning in Combinatorial Auctions: Computing Approximate Bayesian Nash Equilibria via Pseudogradient Dynamics 
S. Heidekrüger, P. Sutterer, N. Kohring, M. Fichtl, and M. Bichler
presented at the 2021 AAAI Workshop on Reinforcement Learning in Games (AAAI-RLG-21) [pdf],
and at the 2020 Workshop on Information Technology and Systems (WITS20)

Multiagent Learning for Equilibrium Computation in Auction Markets,
S. Heidekrüger, P. Sutterer, N. Kohring, and M. Bichler,
AAAI Spring Symposium on Challenges and Opportunities for Multi-Agent Reinforcement Learning (COMARL-21), March 2021 [pdf ]

Learning bayesian nash equilibria in auction games.
S. Heidekrüger, P. Sutterer, N. Kohring, and M. Bichler.
INFORMS Workshop on Data Science, Online, 2020.

Computing approximate Bayes-Nash Equilibria through Neural Self-Play.
S. Heidekrüger, P. Sutterer, and M. Bichler
Workshop on Information Technology and Systems (WITS19), Munich, Germany, 2019.

Teaching

Courses

  • Business Analytics and Machine Learning, Teaching Assistant (Winter Terms 18/19, 19/20, 20/21, 21/22)
  • Seminar "Data Mining", (Summer Terms 2019, 2020, 2021)
  • Seminar "Learning in Games" (Summer Term 2022)
  • Seminar "IT and Management Consulting",  (Winter Term 19/20, 20/21, 21/22)

Completed Student Projects

Emanuel Deisler Smart Beta Investment Strategy Optimization with Deep Reinforcement Learning,
MSc Thesis; Robotics, Cognition, Intelligence (2022)
Manuel Schweitzer Supervised Federated Topic Modelling with User-Generated Labels,
MSc Thesis, Management and Technology (2022)

Calin Buzetelu

Interactive Data Vizualisation for Equilibrium Computation in continuous Bayesian Games,
BSc Thesis, Information systems (2021)
Mitesh Mutha Cooperative Multi-Agent Reinforcement Learning for Train Scheduling,
MSc Thesis, Informatics (2021)
Irina Broda A Game-Theoretic Analysis of Election Campaign Spending,
BSc Thesis, Informatics (2021)
Iheb Belgacem Improving Sample Efficiency in Multiagent Equilibrium learning settings via Advanced Monte-Carlo Methods,
Research Internship, EEIT, (2021)
Daniel Schroter Reinforcement Learning in the MIT Beer Distribution Game,
BSc Thesis, Informatics (2020)
Markus Ewert      Efficient Query Strategies in Preference Elicitation via Deep Learning,
MSc Thesis, Information Systems (2020)
Anne Christopher  Fast Solvers for Batched Constrained Optimization Problems,
MSc Thesis, Mathematics in Data Science (2020)
Lukas Feye  Confidence-Moderated Policy Advice in Multi-Agent Reinforcement Learning,
BSc Thesis, Information Systems (2020)
Florian Ziesche Human Interpretable Machine Learning: A Machine Learning Approach for Risk Scoring,
MSc Thesis, Mgmt & Technology (2019)
Sebastian Rief Detection of anomalies in large-scale accounting data using unsupervised machine learning,
MSc Thesis, Mgmt & Technology (2019)