Yifan Sun’s research is in characterizing and improving optimization methods for today’s computing challenges. Her focus is on machine learning, where the recent explosion of new and exciting complex data and tasks has brought light to many optimization themes, from using fast approximate methods to fitting complex models (like deep neural nets) to learning low-dimensional vector representations. The main goal is to bridge the gap between theory and practice. Past topics have included improved complexity analysis of sparse optimization, better screening techniques for feature selection, decomposition methods for large-scale optimization, regularization via gauges, robust distributed optimization, etc.