Whilst recent times have seen tremendous progress, AI is still far from 'solved'. Machines have only a shallow understanding of the world around them and, to advance to the next stage, they need access to the vast knowledge that humans have about our physical world, culture and emotional behaviour.
To make a machine capable of operating in the world and interacting in a natural manner with us is an exciting grand challenge that will require a range of collaborative advances.
The Centre aims to make next-generation AI systems that have enhanced capabilities in Computer Vision, Machine Reading and Machine Reasoning and Machine Interaction.
To do this we bring world-leading research talent together and work on joint projects that drive progress. Beyond a pure engineering view, we take a principled approach to understanding how to efficiently train machines and interpret their capabilities and performance.
The Centre's approach to creating efficient, effective and responsible AI is inclusive, taking input from computer science, statistical learning theory, probabilistic modelling, engineering, cognitive neuroscience and the social sciences.
Our research groups
The principal goal is to build machines that can read and "understand" unstructured textual information, converting it into interpretable structured knowledge to be leveraged by humans and other machines alike.
Computer vision and image processing aim to extract useful information from images and movies. Possible applications include face recognition, automated analysis of medical images, robotics, visual inspection for production lines and building 3D models of the real world
The group aims to make methodological progress in foundational AI using techniques from statistics, mathematics and computer science.
This group focuses on data-efficient machine learning, probabilistic modeling, and autonomous decision making