In our latest Spotlight Mike Wooldridge, Programme Co-Director for AI, shares his reaction to being awarded this year’s Lovelace Medal and reflects on what kindled his interest in multi-agent systems as an undergraduate. Mike also discusses our upcoming conference ‘AI and data science in the age of Covid-19, of which he is a co-chair.
Firstly, congratulations on being awarded this year’s Lovelace Medal – how does it feel to have won?
Wonderful, obviously, but also somewhat unreal at the moment! The main thing for me is that it’s recognition that multi-agent systems – computer systems that are made up of multiple interacting entities – are now mainstream. This research area has occupied me for my entire working life, and when I started working on it in the late 1980s, agents were just a small and marginal part of AI.
What first sparked your interest in multi-agent systems?
As an undergraduate in the mid-1980s, I specialised in AI and computer networks. This is a decade before the internet went mainstream, but it became obvious to me that networks were the future of computing. And it occurred to me that the future of AI must be the same. Apart from anything else, what makes human intelligence unique is that so much of it derives from our capabilities as social, networked animals – our languages are the most obvious part of that.
If we are ever to build AI systems that have the full range of capabilities as humans, then these systems must have the ability to cooperate, coordinate and negotiate like us. Building AI systems with these social abilities is what multi-agent systems is all about.
You’re a Chair of the Turing’s upcoming ‘AI and data science in the age of COVID-19’ conference. What do you hope to achieve with this event?
There have been many initiatives from the UK’s AI and data science communities in response to the pandemic, coming from individuals, universities and national organisations like learned societies and the NHS. We want to document that activity, and ask what we can learn from it. The main outcome will be a report, which we hope will make recommendations for future research topics and even national policy. As the UK’s national centre for AI and data science, it seems entirely appropriate that the Turing should lead on this.
What would you say is the biggest unsolved problem in AI and data science?
In the past 15 years, one core AI technology – neural networks – has been shown to be incredibly powerful. But frustratingly, we still don’t really understand why and how the technology works. One aspect is the ‘transparency’ problem: a neural network might be very good at something, but it can’t tell you how it does it. A lot of work is going to be needed to overcome this.
When not working, what can you be found doing?
In this stay-at-home year, I’ve enjoyed exploring the Oxfordshire countryside with my wife Janine, and our dog Cyder (the name is a long story!). When the sun shines, there is nowhere else I’d rather be.