Student Stories: Cohort 2 student, Ben Oakes, on tackling training through his first year as a PhD student in Data Science
The Distributed Algorithms CDT (DA CDT) at the University of Liverpool engages its students in impactful, cutting-edge and diverse research, allowing them to collaborate and develop their skills within the university’s data science community. Read my journey with the DA CDT so far…
My Journey
Last year, I graduated from the University of Liverpool with a Masters in Astrophysics. During my degree, I gained not only an appreciation of the space environment, but also for programming and data science. I worked on some small projects surrounding mathematical programming and modelling as part of my degree. This stemmed from a long-routed interest in programming, having taken it at A-level too. Graduating during a pandemic was difficult as it was hard to find available work. Having been knocked back from other PhD opportunities before graduation, I was unsure if it was the right route for me, but when I saw the DA CDT opportunity, I knew I would love it. I applied and was fortunately successful!
I started in October 2020 and have been thoroughly enjoying studying for my PhD ever since.
My Project: Scheduling Surveillance of Space Objects
My project is focused on applying reinforcement learning to sensor management algorithms for satellite surveillance. I am partnered with the Defence Science and Technology Laboratory (Dstl), a branch of the Ministry of Defence. Dstl has an interest in satellite surveillance as we rely on satellites for many parts of our day-to-day lives; they are used for GPS, communication, and internet as well as many other areas. We need to track them with limited sensor availability, especially focusing on possible high-priority events, such as satellite collisions and the resulting debris. Reinforcement learning should help us track satellites over longer time periods with a higher accuracy than we can currently obtain.
My Experience at the DA CDT
Starting a PhD during a pandemic was a nerve-wrecking experience, but the DA CDT has been amazing at getting the new cohort settled in. Being surrounded by students in similar positions with similar focuses, as well as being able to consult with a wide and diverse research group, has been really helpful for getting started and adjusting to life as a PhD student. There are regular CDT and research-specific meetings to ensure everyone is getting on okay and to share progress and interests. The DA CDT also provided lots of training during the early stages of the PhD to help us develop a good background knowledge and the skills we need to complete our research.
Being surrounded by students in similar positions with similar focuses, as well as being able to consult with a wide and diverse research group, has been really helpful for getting started and adjusting to life as a PhD student.
Mini-Project: Parallelising a Gaussian Mixture Model
In the early stages of the PhD, the DA CDT partnered with the STFC Hartree Centre to deliver training on using high-performance computing (HPC) and parallelising code so it can run on multiple processes to speed up computations. Following this training, the students split into teams to parallelise different clustering algorithms – my team focussing on a Gaussian Mixture Model (GMM). Initially, we took time to understand the algorithm and created a standard code. Using the knowledge, we gained from the training course, we parallelised this code to run on the HPC systems at the Hartree centre and attained speed-ups of up to 12x that of the standard code! This proved the use of the training we received, and will allow us to apply this knowledge to our own projects in the future. It was an excellent experience working with the other students, particularly during a time when it was difficult to meet elsewhere (the pubs were unfortunately closed).
I’m looking forward to working with other students in the future and to work on my own project!
Distributed Algorithms CDT
The Distributed Algorithms CDT is an Innovative Data Science, AI and Machine Learning Research Centre, aligning PhD students, academics and industrialists to work together to generate novel solutions to tough data science challenges. If you would like to find out more about our programme and would like to talk about becoming an active member of our CDT community, please visit our website or email kelli.cassidy@liverpool.ac.uk.