Department for the Economy (DfE) funded PhD Studentship in

Department for the Economy (DfE) funded PhD
Studentship in
Machine Learning and Patterning Recognition for chemical and
biological environmental sensor and early warning systems
Applications are invited for a DfE funded PhD studentship tenable in the Faculty of
Computing and Engineering at the Jordanstown Campus (near Belfast). Please note
that a faculty reorganisation is underway at Ulster and this studentship will be based
within the new structure in the Faculty of Computing, Engineering and the Built
Environment.
Project Summary:
The rapid detection of airborne pathogens, the continuous monitoring of hospital
wards to detect bacteria and viruses or even the routine monitoring of atmospheric
and water borne pollutants represent immense scientific and technical challenges of
great importance to society. Complex and expensive laboratory analytical techniques
are required in order to detect biological and chemical species rapidly and accurately.
Current biosensor or physical sensor technology does not offer a realistic potential for
instant detection of these complex entities. In fact many biosensor technologies
struggle to operate reliably and accurately over extended periods in the wild.
We have been working with a new technique for generating analytical chemical data
from gases or liquid droplets in gases, for example from human or animal breath,
industrial gases and aerosols. This technique is called plasma optical emission
spectroscopy (POES); it has shown some interesting results and could be quite low
cost compared to standard laboratory instruments. However the data quality is noisy
and non-stationary and therefore we need to use chemometric and new machine
learning techniques to improve the detection accuracy.
Spectral data from POES contains a wealth of information related to underlying
physical processes but its high dimensionality and high collinearity pose a major
challenge for the development of robust algorithms. To date current mathematical
approaches for handling high dimensionality and high collinearity have not been
explored for POES emission spectra. We will explore new machine learning methods
that are suitable for such open-world classification problem and also methods for
physics-bearing fingerprint extraction. Feature extraction/selection is usually applied
before machine learning is conducted. Since the goal of this project is to understand
the physical composition of the substance, we need to extract/select those features
that characterise the physical composition hence can be used as fingerprint of the
substance for recognition.
This is a collaborative project between physicists and engineers in plasma and sensors
and computer scientists in machine learning and artificial intelligence. We welcome
applications from interested students who have a physical sciences background with
an interest in computation and machine learning or from mathematics or computer
science students with an interest on working on challenging real-world sensor
problems.
Entrance Requirements:
All applicants should hold a first or upper second class honours degree in Physical
Sciences, Computer Science, Mathematics, Engineering or a cognate area.
Applications will be considered on a competitive basis with regard to the candidate’s
qualifications, skills experience and interests. Successful candidates will enrol as of 1
October 2017, on a full-time programme of research studies leading to the award of
the degree of Doctor of Philosophy.
The studentship will comprise fees together with an annual stipend of £14,553 and
will be awarded for a period of up to three years subject to satisfactory progress.
If you wish to discuss your proposal or receive advice on this project please contact:Prof Paul Maguire
[email protected]
http://www.nibec.ulster.ac.uk/staff/pd.maguire
Prof Hui Wang
[email protected]
https://www.ulster.ac.uk/staff/h-wang
Procedure
For more information on applying go to ulster.ac.uk/research
Apply online ulster.ac.uk/applyonline
The closing date for receipt of completed applications is 5 May 2017
Interviews will be held in May 2017