Lovelace lecture 2005

Machines that learn

Speaker: Prof Christopher M Bishop, Assistant Director of Microsoft Research, Cambridge, and Professor of Computer Science at the University of Edinburgh.

Machine learning is a well-established field and plays a central role in areas such as speech recognition, handwriting recognition and face detection.

In the last few years, however, the range of successful applications of machine learning has broadened significantly to include areas as diverse as spam-email filtering, DNA-sequence modelling and computer games.

While improvements in the speed and memory capacity of computers have played an important role in this growth, there have also been some key technical developments.

This talk highlights some of these, such as the use of graphical techniques to represent complex probabilistic models, and the role of approximate inference techniques (such as variational methods) in applying those models.

Recent applications illustrate the talk, which will conclude with some pointers to potential future directions.