MAIN RESEARCH AREAS

• High Performance Computing: The CS theme has deep experience in computing for the most challenging scientific problems. Our research addresses computer arithmetic, architectures, languages, runtime systems, efficient operators and algorithms that push the frontier of performance and scalability. These focus on applications as diverse as large-scale machine learning, health analytics, financial analytics and network analysis on multicore, GPUs or programmable hardware such as Field Programmable Gate Array (FPGA). Our work spans all layers of the design space, including circuit design and synthesis devices and the microarchitecture and chip-level organisation and compute nodes from single-device to server blades and cloud infrastructure. This also extends to emerging computing paradigms, such as quantum and neuromorphic computing. 

• IoT, Cyber-Physical and Autonomous Systems: These systems use smart sensors and lightweight, networked devices to monitor the real world in setting such as power grids, manufacturing, wearables and many others. The CS theme specialize across these areas. We are experts in lightweight, energy-efficient data processing for biosignal analysis in wearables.  We specialize in securing IoT device networks against malicious intruders and smart sensor architectures, combining traditional sensors with custom sensing algorithm and architectures on the sensor. 

• Cloud/Edge: Cloud computing puts vast computing resource and sophisticated data processing pipelines for storage, recall and analysis of data within easy reach. CS has a particular expertise in cloud for advanced manufacturing, healthcare analytics and financial data processing. In other applications where fast response is required, edge computing – computing using devices as the networks edge – offers a highly distributed but massively capable resource. CS has deep expertise in this area. In particular CS has extensively studied the design of dependable computer architectures and hardware design, runtimes for distributed edge platforms, federated learning and algorithms for edge computing. 

FEATURE ARTICLE

J. Lee, D. S. Nikolopoulos and H. Vandierendonck, "Mixed-Precision Kernel Recursive Least Squares," in IEEE Transactions on Neural Networks and Learning Systems, doi: 10.1109/TNNLS.2020.3041677.

OUR IMPACT

• The theme plays a key role in the UK high performance computing ecosystem, as host to one of the six UK government commissioned Tier-II high-performance computing resources, Kelvin-2. It is host to the unique Rakuten Mobile Edge Computing Hub. Its members have also seen significant exposure at the international level; this includes best paper prizes at DATE 2020 for ‘DEFCON: Generating and Detecting Failure-prone Instruction Sequences via Stochastic Search’, was one of only three UK institutions represented at MICRO 2020, where the paper ‘DStress: Automatic Synthesis of DRAM Reliability Stress Viruses using Genetic Algorithms’, was nominated for the best paper prize, and finally were awarded a HiPEAC 2020 Paper Award for the paper ‘HaRMony: Heterogeneous-Reliability Memory and QoS-Aware Energy Management on Virtualized Servers’.  

• Two CS researchers were included in Stanford University’s world top 2% scientists for citation impact. CS houses finalists in the GSMA Mobile World Congress Research Award Competition (2019), UK EPSRC Connected Nations Pioneers (2018), Silver (2018) and Bronze (2017) medallists in the ACM Student Research Competition. We hold roles of significant editorial responsibility including Associate Editorship of IEEE Open Journal of Circuits and Systems (OJCAS) and IEEE Trans. Signal Processing. We held the chair of the IEEE Technical Committee on Design and Implementation of Signal Processing Systems (DISPS), 2020; founding chair of the IEEE Technical Committee on Applied Signal Processing Systems (ASPS), 2021 and seats on the Technical Directions Board of the IEEE Signal Processing Society and Executive Committee (ExCOM) of the IEEE Task Force on Rebooting Computing.  

• It has seen its research included in the EdgeX Foundry open source software project and commercialised via Analytics Engines Ltd.