Dr João Henriques - University of Oxford

Dr João Henriques aims to improve how robots and machines perceive the world through cameras and other sensors and allow safe operation even in unexpected situations. Current artificial intelligence (AI) learning methods need vast quantities of data to understand each new concept, which leaves them vulnerable to blind spots in rare situations. This also creates a perverse incentive for companies to hoard data for AI training, harming privacy. Dr João Henriques is developing AI methods that ‘learn to learn’, adapting quickly with very little data. This will enable more independent and intelligent machines, which are useful for society without needing to upload user data to the cloud.