FST JOURNAL

Summary:

• The ‘edge’ refers to sophisticated devices and sensors available in domestic and work settings which can do a host of things without going back to the core network and data centres
• We've barely got 5G, and 6G is already being explored in a range of industrial settings with focus on being able to communicate to the edge
• Digital Catapult worked with several organisations on the design of Bridge AI which Innovate UK launched aiming to support AI supply in sectors of the UK economy that are currently underserved.

The Digital Catapult is part of the catapult network, which is part of around nine partnership technology and Innovation Centres across the UK. Digital catapult works across advanced technology areas such as 6G future networks, IoT Internet of Things AI, immersive technologies and Quantum. We want to support the adoption, acceleration, and exploitation of new technologies for the benefit of industries, citizens and society. We run accelerated programmes, we undertake research and development, and support take up and adoption of early stage technologies. I have been quite privileged to work in a number of government settings as the director of AI and Ofcom, and in the national infrastructure Commission, as well as at the catapult. Perspectives given here today are my own from across those organisations rather than perspectives of any organisation. 

The edge of the cloud

Technology is moving to the edge, but what do we mean by the edge? We are now very used to being cloud connected and having cloud connected devices. Our phones are essentially connected to the cloud and into data centres at all times. If you are doing a search on Google Maps, using Alexa, using any of those smart assistants, the query that you make is very often relayed back to a data centre over a network, whether it's 4G or Wi Fi, or broadband. The processing that then happens in a data centre comes back out of the network to your device. However, increasingly that interaction is too slow and does not really work in a given situation. It might even be that your data is secure and you do not want to share it back to a data centre or you may need much quicker real time responses or the data may be too big to pass around. The edge refers to a push of computing resources to be close to the user in a network and for sophisticated devices and sensors available in homes, factories, shops, even cars, vehicles, which can do a host of things without going back to the core network and data centres. It is often many sensors and many devices working together. 

Evolution

Both networks and devices are evolving very quickly with ever more sophistication such AI processing on the device and evolving networks. 5G and 6G networks are an important part of this, enabling very high data rates with little delay. For example, you may well have local processing in a factory, a stadium, or a station which enables you to have the processing close to the devices. We have barely got 5G, and 6G is already being explored in a range of industrial settings with focus on being able to communicate to the edge. Examples include deeply immersive technologies such as immersive reality and virtual headsets. In the age of AI where you're passing visual data backwards and forwards to train models, data can be very heavy. Smart factories using robots on private networks within the one factory or connected networks amongst different factories is a good example. You've also got cases like drones and vehicles which need to be connected, and have strong visual processing needs. 

Cutting edge media production

The Nvidia Omniverse platform is an example of a digital twin platform focused on use cases like simulating smart factories and is an example where edge technologies enable cyber physical systems simulation and interaction. AI data preprocessing enables robots, machine lines and cyber physical kind of simulations of factories to be close to what's happening in reality and drive efficiency and communication. There is also a company called Niantic Lightship platform which looks towards the real world, Metaverse, and outdoor immersive experiences. It was the company that bought us Pokemon go a few years ago. Star Wars: The Mandalorian production pioneered advanced virtual production film techniques and Ocado’s warehouse uses smart robots with communications built in to pick up groceries from the floor. All of these are real and are here and now, underpinned by the same building blocks of advanced digital edge technologies, digital infrastructure and human machine interfaces like Virtual Reality and Augmented Reality. 

At Digital Catapult, we have built two advanced media production studios, one in Gateshead in the north of the UK and one in the east of London which are connected. They both have 5G networks, and are connected by very low latency networks which enables cutting edge media production for things like generative AI to be adopted. We put them together in the pandemic when you could not fly actors over from the US, so it enables you to do things like explore new approaches for example when you may have actors in different geographic locations but working on the same digital set. We also hosted a 5G festival which involved a pop band with a singer at the 02 Arena in London, a guitarist in Metropolis Music Studio and drums and other instruments down at the Brighton Dome; all of them connected by very low latency networks. They were able to perform as a live band, seeing each other with immersive headsets.

Bridge AI

There is a programme that we are working on in partnership with several companies here including Innovate UK, the Standards Institute and the Hartree centre called Bridge AI. It is a programme that will enable AI to be adopted and to flourish in some quite different areas to the UK’s usual avenues. Only 5% of AI startups in the UK focus on traditional sectors such as transport, manufacturing, agriculture and construction. The large majority of them are focused on data science consultancies. To address this situation, Digital Catapult worked closely with Innovate UK and the relevant UK funding agency to design and launch Bridge AI. It aims to support AI supply in sectors of the UK economy that are currently underserved. It’s quite a rich and dynamic programme with a lot of organisations around the UK involved. To pick a couple out of a hat, we have VeuNex Global an oil and gas company specialising in tackling health and safety in that very dangerous environment using AI and computer vision to make things safer for people to work. We also have Neural Echo Labs, which is a UK based computer games startup looking at immersive virtual experiences through brainwave communications, blending brain computer interfaces, VR and AI together. Through the Bridge AI programme, a colleague from the Information Commission became a mentor of that startup, and provides ongoing engagement and support. 

The rate of progress of AI is exemplified by a I recent example shown on X (formally Twitter) attempting to reconstruct what humans imagine from MRI data. It was an example of functional MRI imagery where people are shown an image of say a giraffe which in an MRI, and then AI is able to take the data from the functional MRI, and construct what the person is looking at. It just shows some of the sorts of examples that we may expect to run into here.