| Bio | Nando de Freitas is an associate professor in Computer Science and Cognitive Systems at the University of British Columbia (UBC). He is also an associate member of the Statistics Department. He obtained his PhD on "Bayesian methods for neural networks" at Cambridge University in 2000. During the next two years he was a postdoctoral scholar at UC Berkeley with Prof Stuart Russell. Nando has contributed to many areas of machine learning including fast n-body methods, hierarchical Bayesian models for model selection, Markov decision processes, Bayesian optimization, robotics, image tracking, Bayesian experimental design, multimedia retrieval and machine translation approaches for object recognition.
Monte Carlo methods have been central to Nando's research. In 2000, he edited the first book on sequential Monte Carlo methods with Arnaud Doucet and Neil Gordon. This booked played an important role in unifying the field and it is now the 5th most cited computer science publication in 2000 according to the Citeseer index. With his students and collaborators, Nando has contributed to the development of many well known Monte Carlo methods, including Rao-Blackwellized particle filters for dynamic Bayesian networks, unscented and boosted particle filters for tracking, efficient marginal particle filters and particle smoothers, sequential Monte Carlo methods for Boltzmann machines (hot coupling), tree sampling methods for random fields and reversible jump MCMC algorithms for control and planning.
|