Deep reinforced active learning

In this talk I will give an overview of my research at GSK .ai. I introduce a paper in the field of active learning and the application of deep reinforcement learning to image classification in the active learning domain. In this work we aimed to solve several major biases or oversights which exist in the literature in active learning, a machine learning technique widely used in many industrial applications. The paper I will present provides initial solutions to the aforementioned issues in the field, applied to medical imaging datasets. My group members are now developing the technique for much wider-spread applications across the domain of deep learning in drug discovery.

• Speaker: Emma Slade – GSK
• Wednesday 30 November 2022, 14:00–15:00
• Venue: Centre for Mathematical Sciences, MR4.
• Series: CCIMI Seminars; organiser: Paula Smith.