Uncertainty Quantification for Robust AI through Optimal Transport
In this project, we will develop a set of tools centred on the notion of optimal transport theory to quantify uncertainty of the machine learning models, allowing the models to understand what they do not know.
Start date1 July 2021
- Full UK/EU tuition fee covered
- Stipend at £15,285 p.a.
- RTSG of £1,000 p.a.
- Personal Computer (provided by the department)
Funding sourceThe University of Surrey, Project-led Studentship Award.
Machine learning has been making decisions that affect our lives. Yet, we often cannot even tell whether they are uncertain about their decisions. In this project, we will develop Bayesian techniques with tools from the optimal transport theory to better represent and quantify uncertainties in machine learning models. While theoretical results are promising, the deployment of the optimal transport theory in a wide range of machine learning applications is limited due to its heavy computational burden. We will derive algorithms for uncertainty propagation and quantification based on computationally efficient approximate optimal transport methods. The resulted toolkit will be validated on a real-world clinical application and is transferable across a wide range of safety-critical AI applications.
The successful applicant will be supervised by Dr Yunpeng Li and co-supervised by Professor Wenwu Wang. The PhD student will be based at the Nature Inspired Computing and Engineering (NICE) research group in the Department of Computer Science at the University of Surrey. The student will also benefit from resources from the Centre for Vision, Speech and Signal Processing in the Department of Electrical and Electronic Engineering at the University of Surrey.
Related linksNature Inspired Computing and Engineering (NICE) research group Department of Computer Science Centre for Vision, Speech and Signal Processing
A Bachelor’s degree or above in Computer Science, Electrical Engineering, Statistics, Mathematics, Physics or similar (a First Class or good Upper Second Class Honours degree, or the equivalent from an overseas university).
This studentship is for UK, EU and overseas students. Overseas (non-UK/EU) students will need to fund the international tuition fee difference themselves.
IELTS requirements: If English is not your first language, you will be required to have an IELTS Academic of 6.5 or above (or equivalent), with no sub-test score below 6.
How to apply
Please click ‘Apply’ on the Computer Science PhD page.
Please prepare to submit your CV; degree certificates and transcripts; names of 2 referees (ideally uploading 2 references at time of application also); and research proposal (including examples of previous project work).