Freddie Bickford Smith

I'm a DPhil student in the AIMS CDT and a Kellogg College scholar at the University of Oxford. My aim is to enable computers to perform well in an increasingly broad range of tasks. Machine learning, in which a computer develops ability through experience, seems to be the most promising way of achieving this. I hope to contribute to advances in machine learning both by suggesting new techniques and by developing scientific understanding of existing ones.

My most recent research looked at the influence of attention, the ability to focus on relevant information, in visual perception. This began as an MSc project at University College London, supervised by Brad Love, Brett Roads and Ed Grefenstette. I revised and condensed my MSc thesis to form a short paper published at 'Bridging AI & Cognitive Science' (ICLR 2020). A more advanced version of this project is in progress.

Previously I studied mechanical engineering at the University of Bristol. I worked with Fabrizio Scarpa for my MEng thesis on ways of connecting components made of carbon-fibre-reinforced materials, and with Ben Hicks for my BEng thesis on using 3D printing to make a crash structure for a car.