Ayyüce Begüm Bektaş, PhD

Machine Learning | Optimization | Computational Biology

My research lies at the intersection of the mathematical foundations of machine learning, optimization, and statistics, with a focus on algorithms that are both theoretically rigorous and computationally efficient for biological applications. I study how feature representations and similarity measures shape model performance, and I design principled learning frameworks that reveal patterns and dependencies in high dimensional, highly correlated data. By integrating tools from optimization, probabilistic modeling, and statistical learning, I develop methods with formal guarantees on generalization, convergence, and robustness. Motivated by real world problems in healthcare, particularly cancer research, I collaborate across disciplines to translate these methods into practice. My work builds on a strong background in mathematical modeling and large scale optimization, and on doctoral studies that emphasized kernel methods, regularization, and model interpretability. I now seek to extend this foundation to create transparent, reliable, and impactful machine learning solutions for complex biological systems.