The UK government is backing a multi-million-pound initiative to help researchers improve artificial intelligence so it can be used more effectively in key areas such as healthcare.
The research program, entitled Machine Learning for Tomorrow: Efficient, Flexible, Robust and Automated, has been launched by Microsoft and the University of Cambridge to try to make AI smarter, easier to create and useful to more people.
The Government will be adding to Microsoft’s funding, supporting the university as part of its Prosperity Partnerships scheme. This initiative offers funding to universities to leverage industry commitments in a range of areas, including healthcare, technology and manufacturing. Microsoft will not receive any of the Government money.
The Microsoft-University of Cambridge project will last for five years and seek to find better ways of compiling and using the data that forms the basis of AI, simplify the development of AI models and reduce errors and bias in real-world applications of the technology.
AI is revolutionising industries across the world, and the Government has identified AI has one the UK’s four Grand Challenges. PwC found that that AI could help the UK economy grow GDP by up to 10.3%, while spending power could increase by between £1,800 and £2,300 per household.
The technology is embedded in many of Microsoft’s most popular programs, such a PowerPoint, Word and Outlook. It has potential in healthcare to help doctors spend more time with patients and reduce delays; help designers build better gaming experiences, and improve how staff communicate and work in businesses across the world.
However, AI has limitations. The technology typically requires a lot of data, which needs to be free of errors and arranged correctly. This is usually completed manually, with researchers spending huge amounts of time preparing and labelling hundreds, thousands or sometimes millions of pieces of information so computer systems can understand it. If this data is comprised of images, these must be clear and sharp or the AI may misinterpret the picture, which can lead to errors or bias. Once created, AI is good at the task it has been assigned to do but is often poor at other tasks, so wide deployment of the technology is limited. Finally, there is a global shortage of people with the right skills who can create AI.
This means that crucial AI systems can’t be built by many people and organisations yet. Microsoft and the University of Cambridge hope to solve some of these issues with their research, so AI can be developed more widely in the UK and help people across the world.
Microsoft announced the tie-up with the University of Cambridge at its Future Decoded event last year as part of a plan to boost the number of AI researchers in the UK.
