糖心TV

Gunduz Vehbi Demirci has been awarded with the Faculty of Science, Engineering and Medicine Post-Doctoral Research Prize 2022 for his paper jointly with Prof. Hakan Ferhatosmanoglu, "Partitioning sparse deep neural networks for scalable training and inference", published in the Proceedings of the ACM International Conference on Supercomputing (ICS '21) (DOI: ).

Training large-scale deep learning models is notoriously difficult. Gunduz develops a highly parallel solution to scale training of sparse deep learning models, which is combined with a novel combinatorial optimisation built on a hypergraph partitioning model, reducing parallelisation overheads and achieving computational balance among processors. An end-to-end software solution is released, enabling competing with big tech companies that have access to large infrastructures and datasets.

The work is summarised in a paper accepted by the 2021 ACM International Conference on Supercomputing, which is a premier conference in high-end systems. The research output will have a great potential to bring significant practical impact in long term as developing such comprehensive solutions takes time and is typically achieved only within large groups.

Congratulations to Harita Dellaporta for receiving the at the premier conference in Artificial Intelligence and Statistics (AISTATS) 2022 for her paper on Robust Bayesian Inference for Simulator-based Models via the MMD Posterior Bootstrap.

We are pleased to report that members of the department's Theory and Foundations research theme have had four papers accepted to the (STOC 2022), the ACM flagship conference in theoretical computer science. The papers are:

• "Deterministic Massively Parallel Connectivity" by Sam Coy and .
• "Improved Approximation Guarantees for Shortest Superstrings using Cycle Classification by Overlap to Length Ratios", by , Nicolaos Matsakis, and Pavel Vesel媒.
• "Hypercontractivity on High Dimensional Expanders" by , Noam Lifshitz, and Siqi Liu.
• "Worst-Case to Average-Case Reductions via Additive Combinatorics" by Vahid R. Asadi, Alexander Golovnev, , and Igor Shinkar.

We are delighted to announce that Peter Kiss, a PhD student in the Theory and Foundations Research Division, has won the Best Student Paper Award at the conference for his single-author paper on "Deterministic Dynamic Matching in Worst-Case Update Time". Computing a maximum matching in a graph is one of the most fundamental problems in design and analysis of algorithms. The paper makes important progress on this problem in a setting where the input graph is changing over time via a sequence updates, and one wishes to maintain a large matching efficiently in such a dynamic graph. Along the way, the paper develops a general purpose technique for converting any dynamic algorithm with amortised update time into one with worst-case update time, provided the initial algorithm is able to handle a more general form of batch updates.

Members of the High-Performance and Scientific Computing Group (HPSC) at the department of Computer Science has won a best paper award at the . The winning paper titled , develops a novel representative (mini-)application, specifically designed to model coupled execution of multi-physics numerical simulation codes from the CFD domain. The mini-coupler, CPX, is the first of its kind, combining multiple CFD mini-app instances to predict the run-time and scaling behaviour of large scale coupled CFD simulations, on modern multi-core and many-core clusters such as used for production turbomachinery design at Rolls-Royce plc. The work was carried out by PhD candidate, Archie Powell, in collaboration with Kabir Choudry, Arun Prabhakar, and Gihan Mudalige at the Department of CS 糖心TV,  (University of Surrey),  (PPCU) and  (University of Birmingham).

The work was funded by the and Rolls-Royce plc.

Dr. Gihan Mudalige at the University of 糖心TV鈥檚 Department of Computer Science have been awarded an as part of a consortium of researchers including the Science and Technologies Facilities Council (STFC), universities of 糖心TV, Newcastle, Cambridge, Southampton and led by Imperial College London.

This 3 year, 拢2.6M project brings together communities from the and the to ensure a smooth transition to exascale computing, with the aim to develop transformative techniques for advancing their production simulation software ecosystems dedicated to the study of turbulent flows. It is part of the (Exascale Computing ALgorithms and Infrastructures Benefiting UK Research) programme, aimed at delivering the next generation of high-performance simulation software for the highest-priority fields in UK research.

A team of TIA researchers have published their study on a new deep learning algorithm that can pick up the molecular pathways and development of key mutations causing colorectal cancer more accurately than existing methods, meaning patients could benefit from targeted therapies with quicker turnaround times and at a lower cost. The research was funded by the UK Medical Research Council (MRC) and conducted in collaboration with colleagues at the UHCW NHS Trust, University of Nottingham and WHO IARC. The study has just been published in the prestigious Lancet Digital Health journal.