Robust Low-Dimensional Representations for Noisy Real-World Data
About the Role
Real world datasets are often plagued by label and input noise due to variability in data collection, annotation errors, and incomplete records. Fully curating such datasets is costly and impractical, and models trained only on idealised or clean data often fail to generalise to other tests sets such in deployment. This project develops noise-resilient AI models by jointly learning low-dimensional representations for both the data and model parameters within the training phase to build models that learn from both aleatoric and epistemic uncertainty and become robust and generalisable. The project is in close collaboration with the National Physical Laboratory and benefits from the scientific environment and resources provided by the Centre for Vision, Speech and Signal Processing CVSSP and the Institute for People-Centred AI at the University of Surrey.
A generous stipend is offered in addition to funding for UK-level tuition fees and research training.
We are looking for applicants with a degree in Computer Science, Mathematics, Physics, or Engineering. Prior experience in AI is necessary. Prior experience in tomographic imaging and medical physics would be advantageous but is not required.
About You
The scholarship is open only to candidates that qualify to pay UK-level (home) tuition fees. A First Class undergraduate degree or MSc with Distinction (or equivalent overseas qualification) in mathematics, computer science, physics or engineering. Excellent mathematical, analytic, and programming skills Previous experience in AI Previous experience in tomographic imaging would be advantageous.
Link to application - Robust Low-Dimensional Representations for Noisy Real-World Data at University of Surrey on FindAPhD.com
For further information, please contact Spencer Thomas, Principal Scientist at NPL.
About Us
The National Physical Laboratory (NPL) is a world-leading centre of excellence that provides cutting-edge measurement science, engineering and technology to underpin prosperity and quality of life in the UK. Find out more about what it is like working here - The measure of us - Overview
NPL and DSIT have strong commitments to diversity and equality of opportunity, and welcome applications from candidates irrespective of their background, gender, race, sexual orientation, religion, or age, providing they meet the required criteria. Applications from women, disabled and black, Asian and minority ethnic candidates in particular are encouraged. All disabled candidates (as defined by the Equality Act 2010) who satisfy the minimum criteria for the role will be guaranteed an interview under the Disability Confident Scheme. At NPL, we believe our success is a result of the diversity and talent of our people. We strive to nurture and respect individuals to ensure everyone feels valued by treating everyone on the basis of their own individual merits and abilities regardless of their own or perceived identity, as part of our commitment to diversity & inclusion, we ensure we’re creating an environment where all our colleagues feel supported and welcome. More about this on our Diversity & Inclusion page. We are committed to the health and well-being of our employees. Flexible working and social activities are embedded in our culture to create a positive work-life balance, along with a broad range of rewards, benefits and recognition . Our values are at the heart of what we do, and they shape the way we interact, develop our people and celebrate success. To ensure everyone has an equal chance, we’re always willing to make reasonable adjustments to the recruitment process. If you would like to discuss, please contact us.
Recommended Jobs
Financial Controller- Construction
You will be reporting into the Operations Director on a daily basis and the duties of the role will include but not limited to: Manage monthly P&L balance sheet and cash flow forecast P…
Content Marketing Specialist at SetSales
Do you thrive in a fast-paced environment, where you gain a tremendous amount of responsibility? Do you want to be part of an exceptional entrepreneurial team with young, motivated and extremely driv…
Drama ECT - Independent School, Camden
A prestigious Independent School in Camden requires a permanent, full-time Early Career Teacher (ECT) of Drama, starting January 2026. Start your career in a vibrant Arts department. About the S…
Data Officer - Superb Secondary School - Sutton - ASAP Start
Are you a meticulous, organised, and proactive professional seeking a rewarding role in an educational environment? We are recruiting a full-time Data Officer to join a prestigious Secondary school…
Recruitment Researcher - Tax
Up to £30,000 + commission - London/ home working Are you an experienced Recruitment Researcher/Resourcer looking for your next role? Are you looking for a truly supportive agency where you can …
Platform Security Engineer (DevSecOps) - Contract
Platform Security Engineer (DevSecOps) – Contract Qube Research & Technologies (QRT) is a technology and data-driven quantitative investment manager. We build systems that directly support resear…
RVN Required for Fixed Term Maternity Contract in Middlesex
VetPro Recruitment are searching for an RVN for a fixed term 12 month maternity contract in a busy hospital in Middlesex. The role would suit a newly qualified or soon to be qualified RVN who enjoys…
KS2 Teacher | Dagenham | January 2026
Are you a versatile KS2 Teacher looking to teach across Key Stage 2 in a supportive Dagenham school from January 2026? Would you like to join a setting with strong subject leadership and a focus on e…
Reception Teacher — Kingston
Are you an enthusiastic Reception Teacher looking for a Full-Time role from January 2026? A warm, Good-rated primary in Kingston seeks a skilled Reception Teacher to join its Early Years phase on a F…
Head of Governance (FTC)
Salary: £55,000 per annum (FTE), pro rata Hours: 21 hours per week (0.6 FTE) Location: London Coliseum We’re looking for a mission-driven Head of Governance to help shape and strengthen go…