Postgraduate Studentships

Proposed Project Title: Signal Processing for Big Data
Principal Supervisor: Dimitrios Nikolopoulos and Michail Matthaiou

Project Description:

The information explosion caused by the seamless expansion of smart devices, the Internet and e-commerce platforms, has triggered an intensive amount of interest into analysing huge data. Yet, according to Google's chief economist ''Data are widely available, what is scarce is the ability to extract wisdom from them''. In this context, a critical challenge with huge datasets is outlier detection. Generally speaking, outliers refer to observations that do not conform to the expected patterns in high-dimensional datasets. Therefore, it is of paramount importance to develop mechanisms that can identify outliers robustly, effectively and with low-complexity. Capitalising on the combined expertises of the two Clusters on theoretical signal processing and big data sets, algorithmic design and high-performance computing, the successful PhD candidate will develop novel schemes to detect outliers in big datasets.

Objectives:

The challenges that this project will address are multifaceted and can be organised in the following work packages:

• Modelling of outliers in huge outlying datastreams, where a few datasets are significantly different from the other data set constituents.
• Instead of rendering disturbance-free data, which induces a prohibitive complexity burden, we will seek to develop novel data-adaptive outlier detection schemes based on different detection objectives (labelling vs sorting technique) and available information (supervised vs universal approaches)
• Dimensionality/Complexity reduction using advanced signal processing techniques: randomised regression, group sampling and adaptive sampling
• Scalable and scale-free (elastic) implementation of data processing algorithms over disaggregated storage, using new techniques for virtualisation and workload distribution between mobile clients and datacentres

Academic Requirements:

2:1 or higher degree in Electrical Engineering, Computer Engineering, Computer Science, and Mathematics. A strong mathematical background is highly desirable.  English Language Testing System (IELTS) 6.0 with a minimum of 5.5 in all four elements of the test or equivalent. 

GENERAL INFORMATION

The latest starting date of this project is 1 October 2015 with an anticipated duration of 36 months. The stipend of the successful candidate will be approximately £15,000 per annum. The deadline for submission of applications is 28 February 2015. The successful candidate should have a strong mathematical background in signal processing, linear algebra, optimization and regression techniques. Moreover, the candidate should have relevant research experience in software for large data sets, to include applications, libraries, system software, or tools.

This studentship is co-funded by the European Commission under grant agreement H2020-644312 (Project RAPID).

Applicants should apply by 28 February 2015 through the Queen’s online application portal at:https://dap.qub.ac.uk/portal/

Further information available at:

http://www.qub.ac.uk/schools/eeecs/StudyattheSchool/PhDProgrammes/

Contact detail

Supervisor Name: Dr Michail Matthaiou
Tel: +44 (0)28 9097 1702

QUB Address:

The Institute of Electronics, Communications and Information Technology
Queen’s University Belfast
NI Science Park
Queen’s Road
Queen’s Island
Belfast
BT3 9DT

Email: m.matthaiou<στο>qub.ac.uk

Supervisor Name: Professor Dimitrios S. Nikolopoulos

QUB Address:

Computer Science at Elmwood
ECS1 Elmwood Avenue
BT9 6AZ
Belfast, UK                                                          

Email: d.nikolopoulos<στο>qub.ac.uk

Deadline for submission of applications is 28 February 2015.

For further information on the Research Area click on link below:

http://www.qub.ac.uk/research-centres/HPDC/

http://www.ecit.qub.ac.uk/Research/WirelessCommunicationSystems/