Mathematician · PhD
Yunus Emre Demirci
I am a Postdoctoral Assistant Professor at the University of Michigan. I completed my PhD in the Department of Mathematics and Statistics at Queen’s University, under the supervision of Professor Serdar Yüksel. I hold an MSc in Mathematics (2021, advised by Prof. Ümit Işlak) and a BSc in Computer Engineering (2019), both from Boğaziçi University.
My research lies in stochastic control, nonlinear filtering, and partially observed Markov decision processes (POMDPs), with a focus on the robustness, stability, and approximation of control and filtering systems under partial information. More recently, I have been working on continuous-time stochastic systems under partial observation, especially diffusion processes.
Research Interests
- Stochastic Control Robustness and stability of stochastic control in complex systems.
- Machine Learning Reinforcement learning theory, with a focus on large-scale systems and challenges associated with partial information.
- Probability Theory and Applications Ergodicity and stability of non-linear filtering processes.
- Graph Theory Extremal graph theory and cover times in Markov chains.
Publications
Preprints
- Demirci, Y. E., Kara, A. D., and Yüksel, S. Sensitivity of Filter Kernels and Robustness to Incorrect Transition and Measurement Kernel Perturbations in Partially Observable Stochastic Control. arXiv:2508.10658v3
- Demirci, Y. E., Kara, A. D., and Yüksel, S. Refined Bounds on Near Optimality Finite Window Policies in POMDPs and Their Reinforcement Learning. arXiv:2409.04351
Journal Articles
- Demirci, Y. E., and Yüksel, S. (2025). Unique Ergodicity of Non-Linear Filters via Reachability and Uniform Weak Continuity. Electronic Communications in Probability, Vol. 30, paper no. 82, 1–15. doi:10.1214/25-ECP732
- Demirci, Y. E., Kara, A. D., and Yüksel, S. (2024). Average Cost Optimality of Partially Observed MDPs: Contraction of Non-Linear Filters and Existence of Optimal Solutions and Approximations. SIAM Journal on Control and Optimization, Vol. 62, Iss. 6, 2859–2883. doi:10.1137/24M1643736
- Demirci, Y., Işlak, Ü., and Özdemir, A. (2023). Mixing time bounds for edge flipping on regular graphs. Journal of Applied Probability, 1–16. doi:10.1017/jpr.2023.3
- Çiçeksiz, A., Demirci, Y. E., and Işlak, Ü. (2023). The Variance and the Asymptotic Distribution of the Length of Longest k-alternating Subsequences. Discrete Mathematics and Theoretical Computer Science, 25(1). doi:10.46298/dmtcs.10296
- Demirci, Y. E., Ekim, T., and Yıldız, M. A. (2023). Defective Ramsey Numbers and Defective Cocolorings in Some Subclasses of Perfect Graphs. Graphs and Combinatorics, 39(1), [18]. doi:10.1007/s00373-023-02612-4
- Demirci, Y. E., Işlak, Ü., and Yıldız, M. A. (2023). On a cover time problem on a dynamic graph with steps at random times. Statistics and Probability Letters, 197, [109821]. doi:10.1016/j.spl.2023.109821
- Demirci, Y. E., Ekim, T., Gimbel, J., and Yıldız, M. A. (2021). Exact values of defective Ramsey numbers in graph classes. Discrete Optimization, 42, [100673]. doi:10.1016/j.disopt.2021.100673
Conference Proceedings
- Demirci, Y. E., Kara, A. D., and Yüksel, S. (2025). Sensitivity of Filter Kernels and Robustness to Incorrect Transition and Measurement Kernel Perturbations in Partially Observable Stochastic Control. IEEE 64th Conference on Decision and Control (CDC 2025), Rio de Janeiro, Brazil, December 10–12, 2025.
- Demirci, Y. E., Kara, A. D., and Yüksel, S. (2025). Wasserstein Regularity of Nonlinear Filters As Belief-MDPs, and Implications on Ergodicity, Optimality and Learning for POMDPs. American Control Conference (ACC 2025), Denver, CO, USA, July 8–10, 2025.
- Demirci, Y. E., Kara, A. D., and Yüksel, S. (2024). Average Cost Optimality of Partially Observed MDPs: Contraction of Non-Linear Filters and Existence of Optimal Solutions and Approximations. American Control Conference (ACC 2024), Toronto, Canada, July 8–12, 2024.
Selected Talks
- Nov 2025 Stability, Approximation, and Robustness of Optimal Policies in POMDPs. ISS Informal Systems Seminar, McGill University, Montréal, QC, Canada.
- Jul 2025 Wasserstein Regularity of Nonlinear Filters As Belief-MDPs, and Implications on Ergodicity, Optimality and Learning for POMDPs. ACC 2025, Denver, Colorado, USA.
- May 2025 Stability, Ergodicity, and Optimal Control in Partially Observed Markov Processes: Wasserstein Contraction Approach. 10th Meeting on System and Control Theory, HEC Montréal, QC, Canada.
- Jul 2024 Average Cost Optimality of Partially Observed MDPs: Contraction of Non-Linear Filters and Existence of Optimal Solutions and Approximations. ACC 2024, Toronto, ON, Canada.
- Dec 2023 On Regularity and Ergodicity of Partially Observable Markov Decision Processes. CMS Winter Meeting, Stochastic Control Theory and Applications, Montréal, QC, Canada.
- May 2023 Unique Ergodicity of Non-Linear Filters. 9th Meeting on System and Control Theory, University of Waterloo, ON, Canada.
Full list of talks on the CV page.
Contact
Department of Mathematics and Statistics, Queen’s University.