New UROP opportunities will be listed here for one month, and thereafter will appear on the relevant faculty page (UROP website) until notified otherwise by the relevant member of academic staff.


Title of UROP Opportunity (Research Experience) & DetailsExperienced required (if any)Contact Details and any further Information


Added: 21 February, 2020

UROPs in Centre for Doctoral Training in Nuclear Energy Futures

Various placements across the Faculty of Engineering covering projects across nuclear science and engineering

 The skills and experience required will depend on the project/UROP being offered.

Please view the information and contact details, including the FAQs and the listed opportunities at:

Added: 29 January 2020

Title of Research Experience (Project Title): Towards Computation and Space Efficient Super-Resolution: Algorithms based on Structured Matrices

Project description: Super-Resolution is a signal processing approach promising a resolution much finer than the empirical criterion known as Rayleigh limit. The core idea of super-resolution is to estimate the parameters of sparse signal components where the sparsity is defined in a continuous space. This notion of sparsity over continuum avoids self-interference introduced by the leakage effect of discrete griding and large computational complexity brought by ultra-fine grids. Moreover, recent advantages of super-resolution show that certain super-resolution problems can be formulated as finite dimensional convex optimization problems and hence can be provably solved in polynomial time. 

Further information in final column

Skills and experience required: Linear algebra including eigenvalue decomposition

Contact details: Dr Wei Dai, Room 811, Dept of Electrical and Electronic Engineering, Faculty of Engineering, South Kensington Campus. Email:

Further details about the project: This project targets at tailored algorithms that improve the efficiency of super-resolution in orders of magnitude. There are two technical elements essential to modern super-resolution techniques: one is structured matrices including Hankel and Toeplitz matrices and the other is decompositions of these structured matrices. Literature has shown, for example, that the computational complexity of eigenvalue decomposition or singular value decomposition of Hankel matrix can be reduced from O(N^3) to O(N log(N)) if one switch the general solver to a solver specifically designed for Hankel matrices. This huge improvement of efficiency is paramount in practice. While efficient algorithms for Hankel/Toeplitz matrices have been well studied before, there is no work yet in the literature to explore their potentials to super-resolution induced optimization problems. It is well known that the standard solver of super-resolution starts to become impractical when the dimension of the samples goes beyond hundreds. If successful, this project will provide practical tools to handle super-resolution problems of much larger dimension. This project helps student obtain deep understanding of linear algebra and numerical computations, acquire practical skills of numerical programming and managing related projects, gain much insight in time series related data processing, earn experiences in optimization techniques, and catch glimpses of modern super-resolution theory and techniques.

Enter table footer here
Summary of the table's contents