Speaker:Prof Zoubin Ghahramani (Dept. of Engineering)
Venue: Winstanley Lecture Theatre
Time: 24/10/2011 20:30, drinks from 20:15

Information plays a central role in 21st century science, commerce and society. We have huge data sets of measurements collected from large-scale scientific experiments, exciting commercial opportunities arising from exploiting web-scale information, and vast stores of knowledge available to society on the internet. Probabilistic approaches for modelling uncertainty and learning from data are essential to the effective use of these vast stores of information. Modern probabilistic approaches to building learning machines are grounded in the mathematics of the 18th century Reverend Thomas Bayes. I will describe the foundations of this field and our recent work on stochastic processes and nonparametric statistics, along with examples of a number of applications to big data problems such as information retrieval, recommendation, genomic data analysis, financial prediction, and robotics.