My research interests include machine learning, statistical learning theory, and adaptive control, in particular with a focus on statistical methods based on convex optimization, kernel methods, boosting methods, semi-supervised learning, structured classification, and reinforcement learning.
Peter Bickel's research spans a number of areas. In his work on semiparametric models, he uses asymptotic theory to guide development and assessment of such models. His studies of hidden Markov models, which are important in such diverse fields as speech recognition and molecular biology, are directed toward understanding how well the method of maximum likelihood performs. He is also interested in the bootstrap, in developing empirical statistical models for genomic sequences. He is a co-author of the well known book Mathematical Statistics: Basic Ideas and Selected Topics.
My research focuses on probabilistic graphical models, kernel machines, nonparametric Bayesian methods and applications to problems in bioinformatics, information retrieval, and signal processing.
My research interests include graphical models applications to signal processing and communication; distributed statistical inference; as well as information theory and statistics.
My research goal is to solve data problems with real-world impact and at the same time develop new statistical methods to push the frontiers of statistics. My group's current research is driven by solving information technology problems such as those from data networks, remote sensing, neuroscience, and finance, while developing effective statistical or machine learning algorithms (e.g. BLasso) and carrying out related theoretical analysis.