Yusuf Kagan Cicekdag

My name is Yusuf Kagan Cicekdag, and I am currently an undergraduate student at Bogazici University, studying Computer Engineering along with a Double Major in Mathematics. I have previously worked on molecular communication via diffusion. Apart from that, during my Bachelor’s Thesis, I worked on the integration of diffusion policies in reinforcement learning with optimal transport, specifically the Wasserstein Gradient Flow. As part of my thesis, our paper “An Analysis and Improvement Study on the Wasserstein Policy Optimization” (Cicekdag, Y. K., Temir, B. T., Ugur, E.) got accepted at IEEE SIU 2026; where we proposed and compared various improvements for the recently proposed Wasserstein Policy Optimization (Pfau et. al.). During my academic career, I want to focus on the theory of machine/deep learning from a geometric perspective, specifically investigating the relation between the geometric properties of the data manifold and the loss landscape with generalization bounds.