Mustafa Mert Çelikok

Department of Mathematics and Computer Science

University of Southern Denmark

Odense, Denmark

I am a tenure-track Assistant Professor at the University of Southern Denmark, in the Data Science Section of the Department of Mathematics and Computer Science (IMADA).

My research focuses on multi-agent reinforcement learning (MARL) and cooperative AI. I study how autonomous agents learn to cooperate with unfamiliar partners, including humans and other AI agents, particularly in complex and changing environments. My broader interests include reinforcement learning, multi-agent learning, continual learning, and human-AI interaction.

Before joining SDU, I was a postdoctoral researcher at Delft University of Technology, where I worked on multi-agent reinforcement learning for collaborative AI systems as part of the Dutch Hybrid Intelligence Centre. I completed my PhD in Computer Science at Aalto University under the supervision of Professor Samuel Kaski, and was a member of the first cohort of the European Laboratory for Learning and Intelligent System’s (ELLIS) PhD programme. I am currently an ELLIS member.

news

Jun 01, 2026	Our paper “Distributional Active Inference” was accepted to ICML 2026. Also co-organizing the ICLR 2026 Workshop on Multi-Agent Learning and Its Opportunities in the Era of Generative AI.
Jan 01, 2026	Became Co-director of the Danish Pioneer Centre for AI Programme on Learning and Optimal Control in Dynamical Systems (2026–2028).
Sep 01, 2025	Our paper “Value Improved Actor Critic Algorithms” was accepted to NeurIPS 2025.
Sep 01, 2025	Started as a tenure-track Assistant Professor at the University of Southern Denmark, Data Science Section, Department of Mathematics and Computer Science (IMADA).
May 01, 2025	Our paper “On the Complexity of Learning to Cooperate with Populations of Socially Rational Agents” was accepted to AAMAS 2025.

selected publications

Distributional Active Inference

Abdullah Akgül, Gülçin Baykal, M. Haussmann, and 2 more authors

In Forty-third International Conference on Machine Learning, 2026

Bib

@inproceedings{akgul2026distributional,
  title = {Distributional Active Inference},
  author = {Akg{\"u}l, Abdullah and Baykal, G{\"u}l{\c{c}}in and Haussmann, M. and {\c{C}}elikok, Mustafa Mert and Kandemir, M.},
  booktitle = {Forty-third International Conference on Machine Learning},
  year = {2026},
  url = {https://openreview.net/forum?id=m3O4qYQ5hu},
}

The Arbiter Agent: Continually Monitoring Multi-Agent Conversations to Detect Emergent Misalignment

F. Tonini, F. Torrielli, A. D. Lautrup, and 3 more authors

arXiv preprint arXiv:2606.10747, 2026

Bib

@article{tonini2026arbiter,
  title = {The Arbiter Agent: Continually Monitoring Multi-Agent Conversations to Detect Emergent Misalignment},
  author = {Tonini, F. and Torrielli, F. and Lautrup, A. D. and Schneider-Kamp, P. and {\c{C}}elikok, Mustafa Mert and Poech, L. G.},
  journal = {arXiv preprint arXiv:2606.10747},
  year = {2026},
}

On the Complexity of Learning to Cooperate with Populations of Socially Rational Agents

Saptarashmi Bandhopadhyay^*, Mustafa Mert Çelikok^*, and R. Loftin^*

In Proceedings of the 24th International Conference on Autonomous Agents and Multiagent Systems, 2025

Bib

@inproceedings{bandyopadhyay2025complexity,
  title = {On the Complexity of Learning to Cooperate with Populations of Socially Rational Agents},
  author = {Bandhopadhyay, Saptarashmi and {\c{C}}elikok, Mustafa Mert and Loftin, R.},
  booktitle = {Proceedings of the 24th International Conference on Autonomous Agents and Multiagent Systems},
  year = {2025},
}

Value Improved Actor Critic Algorithms

Y. Oren, M. A. Zanger, P. R. Van der Vaart, and 3 more authors

In Advances in Neural Information Processing Systems (NeurIPS), 2025

Bib

@inproceedings{oren2025value,
  title = {Value Improved Actor Critic Algorithms},
  author = {Oren, Y. and Zanger, M. A. and {Van der Vaart}, P. R. and {\c{C}}elikok, Mustafa Mert and Spaan, M. T. J. and Bohmer, W.},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
}

Uncoupled Learning of Differential Stackelberg Equilibria with Commitments

R. Loftin^*, Mustafa Mert Çelikok^*, H. Van Hoof, and 2 more authors

In Proceedings of the 23rd International Conference on Autonomous Agents and Multiagent Systems, 2024

Bib

@inproceedings{loftin2024uncoupled,
  title = {Uncoupled Learning of Differential Stackelberg Equilibria with Commitments},
  author = {Loftin, R. and {\c{C}}elikok, Mustafa Mert and {Van Hoof}, H. and Kaski, Samuel and Oliehoek, F. A.},
  booktitle = {Proceedings of the 23rd International Conference on Autonomous Agents and Multiagent Systems},
  year = {2024},
  pages = {1265--1273},
}

Best-Response Bayesian Reinforcement Learning with Bayes-adaptive POMDPs for Centaurs

Mustafa Mert Çelikok, F. A. Oliehoek, and Samuel Kaski

In Proceedings of the 21st International Conference on Autonomous Agents and Multiagent Systems, 2022

Bib

@inproceedings{celikok2022bestresponse,
  title = {Best-Response Bayesian Reinforcement Learning with Bayes-adaptive POMDPs for Centaurs},
  author = {{\c{C}}elikok, Mustafa Mert and Oliehoek, F. A. and Kaski, Samuel},
  booktitle = {Proceedings of the 21st International Conference on Autonomous Agents and Multiagent Systems},
  pages = {235--243},
  year = {2022},
}