A leading UK particle accelerator scientist has welcomed the UK government’s AI Opportunities Action Plan, announced in January, saying that it could help unlock new frontiers in research such as improving accelerator performance, enhancing particle physics experiments, and driving breakthroughs in various industries.
Managing and processing very large banks of data with AI tools is already impacting accelerator design and operation, enabling breakthroughs such as optimising machine architecture, reconstructing transverse beam distributions using machine learning, and enhancing data analysis in high-rate monitoring systems.
Carsten P Welsch, head of Liverpool's accelerator science cluster based at the Cockcroft Institute and the director of the Liverpool Centre for Doctoral Training for Innovation in Data Intensive Science (LIV.INNO), believes that the UK’s action plan can help propel this further.
“Particle accelerators and light sources rank among the most complex and data-heavy scientific systems in existence,” he said.
“The relationships between machine subsystems are complex and often nonlinear, with system dynamics that change over time across numerous variables and different accelerator components. That is where AI can play a big role,” he said.
The UK’s AI Opportunities Action Plan aims to accelerate the uptake of AI in the public sector, as well as support research and innovation advances and build up the UK’s AI infrastructure, including data centres and supercomputers. The plan contains 50 recommendations for the government, which are now at various stages of implementation.
Liverpool a hub for AI
Part of the plan suggests the creation of “AI Growth Zones” (AIGZs) to more efficiently build data centres catering to the needs of AI. The action plan suggests Culham Science Centre, the UK Atomic Energy Authority’s headquarters, could be chosen as a pilot site for this.
Welsch believes that the wider Liverpool area should be considered as a candidate to become an AIGZ.
“With its specialisms in accelerator science, the creative industries, infection medicine, novel materials, and advanced manufacturing, the Liverpool City Region offers potential for a dynamic innovation ecosystem aimed at solving societal challenges,” Welsch said.
The LIV.INNO centre is already creating a bridge between AI and accelerator physics that could be further built on.
The centre provides researchers with training in advanced AI and machine learning techniques to process, analyse, and optimise data produced by particle accelerators for fundamental physics discoveries.
It also hosts projects on applying data-driven methods to improve accelerator design, tune machine performance, and enhance diagnostic tools as well as developing AI-based reconstruction techniques to improve real-time monitoring and feedback systems in accelerators.
“British scientist Tim Berners-Lee developed his ideas for the World Wide Web while working at CERN. Our students are building on this legacy, pushing the boundaries of AI,” Welsch said. “At LIV.INNO, we are excited to collaborate with industry and government to harness AI's potential, driving innovation and economic growth that will benefit society at large.”
Recently, global IT company Kyndryl announced plans to create a new AI hub in Liverpool.
In response Liverpool City Region mayor Steve Rotheram told the BBC that LCR is “leading the way in the UK’s AI revolution”.
He said: "As a region with a proud history of innovation we're ready to seize the opportunities that AI and digital technology can bring; not just to boost our economy but to improve lives, develop skills, tackle inequality, and ensure no-one is left behind."
LIV.INNO is holding a free online event 'AI for Innovation' on 7 May to provide context for recent AI developments, offer insights from data scientists, and share real-world industry experience from those who have started their journey. An option to select a virtual exhibitor stand to showcase your company and network with others at the conference is also available.
Further information and how to register can be found here.