James Orwell is a Professor of Computer Science with significant industry experience, with contributions to signal processing theory (e.g. CVPR, ICPR) and deploying this research in real-world scenarios (airports, paralympics, saleable products), which has led to a succession of enterprise and knowledge transfer activities, both within and in addition to the computer vision domain. he has created hackathons for student employability & social engagement, and made some media appearances, e.g. BBC Horizon and Radio 4. His first degree was in Physics & Philosophy at Oxford, then a Physics PhD with the Image Processing Group in King`s College London. Thereafter, he`s been a member of Computer Science School (in its various guises) at Kingston University.

Deep Learning and Probability. `Deep Learning` is an exciting development of Neural Networks and Data Science, that has provided breathtaking improvements in performance across a range of detection, recognition and estimation problems. This talk will explore the relationship of deep learning to probability estimates, and how information theory provides the mechanisms for training systems to be increasingly accurate.