Michael I. Jordan

I have a longstanding interest in aspects of statistical inference and decision-making that raise computational and systems problems. One focus of my research has been the area of probabilistic graphical models, which blends probability theory and graph theory to provide statistical models for interdependent collections of variables and accompanying graph-theoretic algorithms for inference. I have developed new graphical model architectures that have had impact in various applied fields, ncluding bioinformatics, computational vision, speech, natural language processing and information retrieval, and I have contributed to the development of a novel framework for inference in graphical models based on variational representations of probability distributions. Another area of focus has been nonparametric inference, including both Bayesian nonparametrics, where I have developed new models based on the area of stochastic processes known as completely random measures, and frequentist nonparametrics, where I have focused on kernel machines, spectral methods, dimension reduction and classification. I have also been interested in the psychophysics of human learning, specifically as manifested in human motor control.