Mu's initial research interest was dimension reduction. In the early years of his faculty career, he devoted much attention to efficient kernel machines for rare target detection and ensemble methods for variable selection. He also worked on algorithms for making personalized recommendations, and applications of machine learning to healthcare informatics.
While ensemble learning continued to captivate his curiosity, in more recent years Mu explored a hodgepodge of different topics—such as evaluation metrics, protein structures, transactional networks, and genetic epistasis. At present, he is studying various problems about dependence modeling, large covariance matrices, and generative neural networks.
Mu received his undergraduate degree from Harvard and his PhD from Stanford. He is an elected Fellow of the American Statistical Association and currently the Director of Data Science at Waterloo.
- Zhu M (2014), "Making personalized recommendations in e-commerce," in Statistics in Action: A Canadian Outlook, J. F. Lawless, Ed., Chapman & Hall, pp. 259 - 268.
- Zhu M (2008), "Kernels and ensembles: Perspectives on statistical learning," The American Statistician, 62, pp. 97 - 109.
- Zhu M (2004), "On the forward and backward algorithms of projection pursuit," The Annals of Statistics, 32, pp. 233 - 244.