© Pint of Science, 2022. All rights reserved.
As we use more and more advanced machines and computers for ever trickier tasks new methods are needed to teach our digital friends what to do. In this series of talks we will be investigating some of the cutting-edge ways that computers can learn that are being researched right here in Leeds.
Reinforcement Learning: Robots Learning by Doing
Dr Matteo Leonetti (Lecturer in the School of Computing, University of Leeds)
Practice makes perfect, we say. One of the most powerful features of the human mind is the ability to learn from experience and generalize to similar situations. Reinforcement learning is a theory which provides the basis for autonomous agents to acquire new behaviours, on their own, by trial and error. The robots of the future will have to be versatile, constantly learn new skills, and adapt to the people with whom they will share the environment. I will introduce you to reinforcement learning, and describe some of my research from video games to robotics.
Deep learning for medical image analysis
Rebecca Stone (PhD student)
Deep learning is the new hot topic in artificial intelligence, and full of potential to impact the medical community, which in turn affects all of us. The thought incites excitement in some, and, understandably, fear in others. How does deep learning work, and what are some of the current successes and challenges in its application to medical image analysis?