Kent academic to give computers human touch

Howard Bowman and Colin Johnson of the Computing Laboratory at the University of Kent at Canterbury (UKC) have been awarded a grant of £150,000 from the Engineering and Physical Sciences Research Council to construct computational models of human attention. The research will be undertaken in collaboration with the Medical Research Council’s Cognition and Brain Sciences Unit in Cambridge, one of the UK’s leading centres for research into human attention.

We live in environments in which many events occur simultaneously and compete for our attention. For example, when standing on a street corner we are subject to a plethora of stimuli: cars passing, conversations amongst pedestrians and street vendors plying their trade. When placed in such environments, humans are very good at prioritising these competing stimuli, directing attention towards the highest priority events and ignoring the rest.

When we perceive a significant event such as a car careering off the road, the current task is interrupted and attention is redirected to responding to the new event by, for example, jumping out of the way of the oncoming danger. In contrast, computer systems do less well: robots struggle to perform effectively in environments in which demands on their processing change unpredictably.

In response to this, the study will be working to construct neural network and logic-based models of how humans direct attention. These models will not only increase our understanding of human attention, they will also allow us to construct computer interfaces that are more sensitive to the human user and to develop robots that perform more effectively in dynamically changing environments.

Howard Bowman is Senior Lecturer in Computer Science at UKC. His two main areas of research interest are computational modelling of human cognition and formal methods. Colin Johnson lectures in Computer Science and his research interests centre on projects that combine computing with biological and mathematical ideas.