Festival of Research keynote lecture: Building a machine that sees

Please note, this page has been archived and is no longer being updated.

Swansea University’s Festival of Research kicks off with a keynote lecture by Microsoft’s Distinguished Scientist and Laboratory Director, Professor Andrew Blake.

 In this lecture, which is free and open to all, Professor Blake will talk about computer science, artificial intelligence and how machines with the gift of sight contribute to our understanding of the nature of intelligence..

Andrew BlakeTitle: Building a machine that sees

‎Speaker: Professor Andrew Blake, Microsoft Distinguished Scientist and Laboratory Director, Microsoft Research, Cambridge.

Date: Monday 25 February

Time: 11 am

Venue: Taliesin Arts Centre, Swansea University

Admission: Admission is free and all are welcome

This event opens Swansea University’s Festival of Research – a week-long showcase of research excellence, supported by Bridging the Gaps, which runs from Monday, February 25 until Friday, March 1, 2013.

There are several examples of machines with the gift of sight, albeit in a primitive way compared to human or animal ability.  For example: systems that can recognise up to 1000 different kinds of objects; face detection in consumer cameras; self-driving cars in California and Oxfordshire; and the Kinect camera, partly developed in Professor Blake’s Cambridge laboratory that recognises gestures, and has sold around 20 million units.

Machine vision systems like these are important in practical settings, but they also contribute to our understanding of the nature of intelligence.

Professor Blake will explore the fascinating question of whether truly intelligent systems will turn out to depend more on theoretical ideas, or simply on learning from data at a sufficiently large scale. We may yearn for theories and models to win the day.  In practice, the highest performing machine systems so far, for vision and also for speech, work with shallow models and intensive training.

Professor Blake will discuss which methodology can ultimately approach the goal of human-scale intelligence - elegant models or big data.