Although it is a rather nebulous term, "Big Data" typically means one of two things: big data in a statistical sense, meaning "Big N" and implying the use of all N data points available, and not just a sub sampling; or "Big" as in "massive data size", implying a stretch of the storage capacity beyond normally available levels - and this is obviously a time dependent meaning as storage capacity is becoming more affordable and plentiful. We have interests in both meanings as part of our Big Data research work within The Digital Centre.
An area of focus within Big Data more generally, is that of Big Data Analytics. Thi subfield focusses more on the analysis and possible deductions that one can make from big data. Our approach to data analytics focusses primarily on graph and network science methods, and deep learning approaches.
A key area for us is use of Deep Learning, Artificial Intelligence algorithms to look for patterns and interesting features in big data sets. Our Deep Learning activities are led by Dr Nina Dethlefs, and make strong use of algorithms and approaches inspired by nature, in a sub field of computational science becoming known as natural computations.
Another area of growing activity is big data associated with sensors and devices on The Internet of Things. The IoT as it has become known, encompasses software and hardware to enable such devices as well as protocols, communications algorithms and of course pattern recognition and data mining for the wealth of data becoming available in this way. Our IoT project activities are led by Dr James Walker.
We are developing software embodying data analytical algorithms for processing very large data sets in a Big N statistical sense using: graph algorithms drawn from our work on network science; deep learning and artificial intelligence techniques to identify and mine data patterns; and immersive augmented and virtual reality techniques for visualising complex data sets.
Alongwith colleagues and collaborators within the University of Hull and in local industry, we are considering how a state of the art massive capacity data centre might look and function. The University operates a state of the art supercomputer - known locally as "Viper" and having 5,500 compute cores (at time of writing). We are presently considering what a separate "big data" supercomputer that is even more focussed upon storage and on big data project work, might look like and how it might be housed in a new custom designed data centre building, so as to best leverage: data protection and cybersecurity issues; processing scalability; storage management; analytic processing; and visual analytic capabilities. We are working on some technologies such as memristors and large scale memristor based memory and storage systems that might enable future data intensive high performance computing systems.
Some of our Big Data projects include:
We were fortunate to be awarded a Cluster of PhD Scholarships - with five new students starting PhDs with us in topics related to Big Data in 2017. There are ongoing opportunities to work with us in the area of Big Data through: contract research and development; collaborative grant funded work; student work placements or student research internships or postgraduate study.
You can contact us via Prof Ken Hawick, Director.