“The UK is doing a fantastic job of research on AI, and a less fantastic job of researching with AI. We’re missing out on a whole industry.”
So said Adrian Wander in the latest episode of Boston Bites.
Christine Foster, CCO at the Alan Turing Institute, which focuses on making sure that the best research makes its way into everyday life, said, “It's well known that the UK is third worldwide in AI. We sit just behind the US and China and ahead of Canada, France, Germany and so forth… there's definitely always a need for a co-ordinated, joined up collaboration in the UK."
Large-scale computing has revolutionised our lives in fields of national importance such as weather and climate modelling, financial services and several large-scale computing (LSC) systems, such as those used for weather forecasting, form part of the UK’s critical national infrastructure.
AI is an essential tool for solving industrial and scientific problems and an important enabler of R&D, as demonstrated remarkably by DeepMind’s recent breakthrough on protein folding. Computing power also underpins key technologies such as Machine Learning and Digital Twins. New opportunities also come with new challenges that the UK must address to realise the true potential of LSC.
Lord Chris Holmes identified that biggest problem the UK faces now is realising how to apply AI, highlighting how the 'applied' piece is flawed – the part that takes all of this potential and turns it into something that grows businesses, transforms people's lives, creates jobs, and energy solutions up and down the country. In order for an intelligent application of AI, Christine suggests an immediate action for Government, saying how "a narrowing of priorities with increased investment in fewer areas" will help drive the UK's knowledge in the AI space.
“If AI is pervasive, the message about AI must also be pervasive”
Dev Tyagi
Although initiatives exist, the UK is lacking government steer. A key point of discussion was the need for a unifying UK strategy for large-scale computing. There is currently not enough compute, not enough data, and there isn’t a sophisticated enough set of AI algorithms being applied to the method to actually learn from the data.
We talk about deep learning and machine learning, but how many people actually really know what that is and how many businesses in the UK understand how to harness the power of that technology? What about the businesses that are being left behind?
Adrian confirmed, “I think we've suffered a bit with piecemeal investments from a number of different research councils. I think we also suffered with Brexit… We need to work together, [and develop] some kind of coherent strategy to cover all of the logical attributes of computing and accelerated computing.”
How can we in the UK differentiate ourselves if we are simply a cookie-cutter and using what everyone else is using? We are a knowledge-based economy and there is a tremendous opportunity here. We can create and nurture the next generation of unicorn companies that will help our UK economy to grow, go out, and sell those UK capabilities across the globe, right after we develop core skills and create the right environment for them to flourish.
What’s more, large-scale computing underpins many of the fields where the UK has comparative strength, including bioinformatics, climate research, materials science, pharmaceuticals, machine learning and fluid dynamics. The UK has the largest software industry in Europe. The UK software industry contributed to direct value-added GDP of £70.3 billion in 2016, and it directly employs nearly 700,000 workers.
Nigel emphasised that driving adoption of AI is fundamental, as is providing compute that drives the adoption of the next generation of AI, adding how “there is a whole bunch of work around AI related to fusion.”
“We've got an opportunity to lead the world on those aspects of healthcare, energy and transportation which are all going to be driven, but this kind of infrastructure is one the government needs to address, because if we don't do it and other countries do, then very quickly, our industries are going to fall behind relative to the other advanced economies.”
Christine elaborated how “it's much about that creative leap and that human ingenuity than it is about any sort of specific technical skill.” She discussed how, in the UK we are: extremely adept in the discovery space and finding new algorithms; how UK AI centres, particularly around research, have strong talent, a great university sector and research institutes; “increasingly so, with more and more companies delivering the products and services to make AI real for everyday life.”
"We're opening up the next phase of compute"
Nigel Toon, Graphcore
AI could be incredibly transformational for UK businesses and the human experience.
“Most could be doing something with their data,” said Adrian, “but they just don't know they can do it.” AI could revolutionise healthcare for example, where the UK has always been a leader.
“Some of the large labs in other countries have seen AI go from a few percent of their compute workload to be close to 50%,” informed Nigel. In light of this, Graphcore developed a next generation IPU processor, BOW, to deliver a huge leap of performance to what is really possible with AI. Nigel believes in "building an ecosystem of capability and skill”, learning from data and making that process as simple as possible, to enable those new innovations and breakthroughs. “We are working together with Boston as one of our key partners to get our systems out to customers and support the next generation of AI across industry.”
The UK needs the infrastructure that provides support to our growing technology community, especially in the AI arena. Where is the advice that they can refer to that allows them to see what is possible? That's what's missing right now. It was concluded that, on a community level, we can all do our part in promoting the resources that are already available.
Young academics want to become entrepreneurs, so we need to create the environment that supports that and allows them to cross that difficult chasm from university to a business.
How do we make sure the research community and scale-up companies have access to the computers necessary? How do we make sure that even the large-scale businesses can compete on a global basis? Currently, people misinterpret that message, and take money out of one pot and put it into another pot for the sake of it, with no long term strategy, but more of a short term experiment.
We need clarification and understanding. Enter Boston Bites.
The panel came together to demystify AI, to amplify how we can use AI “to go beyond what has been possible before and solve complex problems that you couldn't easily write a program to perform,” explained Nigel, “we're here to show people what is possible in their business application".
Boston Bites is a tasteful example of fusing food and thinking.
Our food-for-thought initiative configures new thinking patterns and digestible concepts to fuel further conversation and personal research. Boston Bites mixes up the algorithm of a more traditional panel discussion, fusing together seasoned professionals and hot industry topics -- with that comes a selection of confident and knowledgeable talkers positioned around the table to think, eat, converse and plan to make a difference.
Dev told us that “having Cyrus taking us through building a meal with all of the different flavours that are joining together is very much the same in terms of how you need to take the different ingredients of the compute, the data, the method you're going to use, to train the computer to actually create these AI systems. It's a really helpful environment to demonstrate what we're trying to pull together here.”
A helping hand is required, and that helping hand is Boston - “we’re not just talking about it – we’re helping people do it,” Dev replied. UK focus should shift onto educating all the way from entry-level to university, and bridging the gap between education and the application of technology within organisations.
We must ask ourselves "how do we make sure that those incredible individuals and entrepreneurs are helped along their journey?”
Chris added "we need that connectivity, that collective collaboration, because what we have now, which we've never had in our hands before, are the most powerful, transformational tools.”
The panel discussed some of the most pernicious issues that have dogged humanity for decades: “climate change, health, mobility – AI and high-compute have such a key role to play in all of those,” said Chris firmly. “If we get it right, we could truly get so much closer to the society where everybody can feel that sense of inclusion and empowerment.”
As a round up of events, Chris predicted “when connection, collective collaboration, state, institution, academia, private business all come together to recognise, realise, develop and deploy that AI opportunity in almost every element of everything we do, it will enable us to be the humans and the society that we’ve always believed we could be.”?