We are creating a unified UKRI website that brings together the existing research council, Innovate UK and Research England websites.
If you would like to be involved in its development let us know.

Case study - Alex Marshall

Alex completed a Master’s in Physics at the University of Bristol and is now part of the Centre for Doctoral Training in Data-Intensive Science, supported by UKRI’s Science and Technology Facilities Council. He is part of the Search for Hidden Particles (SHiP) project at CERN which is looking for new particles which may help to explain some of the greatest mysteries in physics, such as how the universe and dark matter work and why the universe is dominated by matter, as opposed to matter and anti-matter being balanced. Alex is using approaches in Machine Learning, a subset of AI, to speed up computational simulations of the experiment and provide improved understanding of the background, increasing the experiment’s sensitivity and increase the chance of detecting new physics.

“I received a significant amount of training in my first year at the CDT and this has proven useful in my overall understanding of the data science. Although Machine Learning is quickly becoming a core area of research in particle physics, most particle physics PhDs are not focused it. Being part of the CDT has allowed me freedom to explore applying Machine Learning ideas to particle physics, and this has become the heart of my PhD which I believe is fairly unique. The freedom I have had has given me a chance to explore what I find interesting, this has made my PhD extremely exciting.

“Machine Learning is becoming increasing prevalent in particle physics, most commonly in signal or background discrimination. We have huge experiments looking for small signals in huge datasets. Being as accurate as possible in understanding and removing background events can increase the power and reach of these expensive experiments.

“Modern particle physics experiments are constantly improving to reach new levels of sensitivity to new physics. Through high luminosities or more accurate modelling of backgrounds for example. Simulation of these environments will also become increasingly computationally expensive, traditional simulation methods may eventually rely on the support of modern faster Machine Learning methods.

“Meeting up with members from across the three universities involved in my CDT has been useful. It has been interesting to hear about their research, especially in poster sessions we have all been involved in. Being in a group of people all using similar techniques or ideas for individual research projects has shown me a wide range of applications and helped me in understanding the potential of these techniques. Being able to share resources has also been helpful, especially at the start when everyone is exploring data science broadly.

“This PhD has given me many opportunities to apply high level Machine Learning techniques to a fairly unique set of physics problems. I have had the freedom to approach these problems from any angle and I have used this to broaden my skill set and understanding. The problems I have faced have been quite specific but they have required a level of wider understanding, from this I believe I am well equipped to tackle Machine Learning problems in whatever career I enter post-PhD. The training provided by the CDT was, rightly, broad and fairly introductory and was useful after a summer away from science. I have used this training as a seed to explore Machine Learning on my own, which was the right approach for me.

“Increasing the level of funding for AI in academia is a key part of the future of the field’s success in the UK. Many students leaving university now are excited to learn and be taught about fields of AI, such as Machine Learning, while still being interested in the physics. Funding Machine Learning PhDs with a science focus in the UK in a is the perfect platform for these students. The general quantitative understanding gained in a physics PhD is not lost whilst the opportunity to explore exciting techniques is there. Solving complex physics problems with Machine Learning can provide more than enough understanding for those leaving academia to thrive in industry. If the focus of this funding is to produce outstanding and highly trained people that can spin out into industry it will be successful.”