Tolga Ok
My goal is to make reinforcement learning (RL) work out of the box. Towards this goal, I develop mathematically grounded methods and insights, and build tools to facilitate RL experiments. I’m currently a PhD candidate at the Delft Center for Systems and Control at TU Delft, supervised by Peyman Mohajerin Esfahani and Mohamad Amin Sharifi Kolarijani.
Research
Value gradients and costates | Oct 2026 | How reward sensitivity travels backward through a stochastic system, and how this differs from REINFORCE.
Software
Publications
Limiting-Kernel Q(λ): Bridging Short and Long Horizons
Tolga Ok, Arman Sharifi Kolarijani, Peyman Mohajerin Esfahani, and Mohamad Amin Sharifi Kolarijani
arXiv 2026under reviewPaperOffline Reinforcement Learning via Inverse Optimization
Ioannis Dimanidis, Tolga Ok, and Peyman Mohajerin Esfahani
Control and Reinforcement Learning through the Lens of Optimization: An Algorithmic Perspective
Tolga Ok, Arman Sharifi Kolarijani, Mohamad Amin Sharifi Kolarijani, and Peyman Mohajerin Esfahani
arXiv 2025under reviewPaperRank-One Modified Value Iteration
Arman Sharifi Kolarijani, Tolga Ok, Peyman Mohajerin Esfahani, and Mohamad Amin Sharifi Kolarijani
Scalable Kernel Inverse Optimization
Youyuan Long, Tolga Ok, Pedro Zattoni Scroccaro, and Peyman Mohajerin Esfahani
Core Skill Decomposition of Complex Wargames with Reinforcement Learning
Kubilay Kağan Kömürcü, Batuhan İnce, Tolga Ok, Emircan Kılıçkaya, and Nazım Kemal Üre
AIAA SciTech 2022Paper