My research interests are broadly in interpretable machine learning, data-driven analytics, and optimization.

Many machine learning problems reduce to solving an optimization problem. My research focuses on developing methods to solve these challenging optimizations problems that arise in many different machine learning settings. In particular, my optimization work has revolved around solving non-convex and non-smooth problems.

Another focus of my research work has been developing interpretable machine learning methods that facilitate the decision-making process for domain experts. Specifically, in my recent work, we developed a hybrid model where an interpretable model is constructed to compete and collaborate with any pre-trained black-box model, gaining model transparency at no or low cost of predictive performance.

I am also interested in data visualization and data-driven analytics with applications in health, sports, and other fields.

Conference Publications

Interpretable Machine Learning

Optimization

Journal Publications

Working Papers

Manuscripts In Progress

  • H. Rafique and Q. Lin. Second-Order Trust-Region Method for Non-Convex Min-Max Problems.