Alina Harbuzova

Alina Harbuzova is a graduate student in electrical engineering and computer science whose work addresses foundational questions at the intersection of statistics, machine learning, and computational complexity. Her research focuses on understanding the computational limits of high-dimensional statistical tasks and developing algorithms with provable guarantees. Alina will use the MathWorks Fellowship to advance two primary areas: establishing average-case equivalence between long-studied statistical models and resolving theoretical challenges in semi-supervised learning on graphs. In both areas, she integrates rigorous analysis with extensive MATLAB-based simulations to guide theoretical insights, test algorithmic hypotheses, and verify model behavior. MATLAB’s linear algebra and regression tools, along with its flexible visualization capabilities, are integral to her workflow. Alina’s work could reshape how researchers define and assess computational difficulty in modern statistics, offering new tools and insights to inform both theory and application across machine learning, data science, and algorithm design.