Manchester Home Dr David C. Wedge













Dr David C. Wedge
Postdoctoral Research Associate
Bioanalytical Sciences Group
School of Chemistry
The University of Manchester
Manchester Interdisciplinary Biocentre
131 Princess Street
Manchester M1 7DN
Dr David Wedge


Personal | Research | Education | Publications | Miscellaneous | Downloads | Contact


Personal

Dr David C. Wedge

School of Chemistry, The University of Manchester, Manchester Interdisciplinary Biocentre, 131 Princess Street, Manchester M1 7DN

Tel: +44 (0)161 306 5145

Email:



Research

Low Cost Sensor Arrays and Genetic Programming
The ability to constantly monitor aqueous and gaseous environments can play an important role in ensuring public safety. Unfortunately, current analytical techniques are often slow, expensive or too specific. This research uses the change in behaviour of semi-conducting materials in response to environmental chemicals at low concentration as an indication of contamination. When arrays of different organic field effect transistors (OFETs) are used as a sensor a multi-dimensional response is induced. This response may then be analysed using statistical or machine learning techniques.

By using Genetic Programming (GP) as a tool for pattern recognition we aim to identify a large range of chemicals. GP is a machine learning technique which uses an evolutionary approach to improve a population of solutions to a problem (computer programs) through the processes of selection, reproduction, mutation and re-combination.

I have written all of the code used to run the GP algorithm from scratch using the Java programming language. All of the code may be downloaded from the downloads section.

General information may be obtained concerning chemical sensors here and concerning Genetic Programming here.

Information on our group's other research in the area of Genetic Programming may be found here.



Education

2005 Manchester Metropolitan University, PhD Computer Science, Wave Overtopping Prediction using Global-Local Artificial Neural Networks. This project aimed to predict the rate of overtopping of sea-walls with a variety of structures and under a variety of environmental conditions. Theoretical adcances were made in the field of hybrid artificial neural networks through the invention of a novel architecture and training method for networks containing both sigmoidal and radial basis function (RBF) neurons. This research took place within the Centre for Mathematical Modelling and Flow Analysis under the guidance of Dr. David Ingram. My thesis is available for download here.

2001 University of Huddersfield, MSc Software Development, Distinction with a prize for the best postgraduate student. This included a research project on the introduction of heuristics into constraint checking during AI Planning under the supervision of Prof Lee McCluskey.

1993 Bretton Hall College, University of Leeds, PGCE Early Years.

1989 Pembroke College, University of Oxford, BA(Hons.), Chemistry 2(i) with Distinction in History and Philosophy of Science

1985 Hewett Comprehensive School, Norwich, 'A' levels in Mathematics (A), Physics (A), Chemistry (A) and English Literature (B)

1983 Southgate Comprehensive School, Enfield, 'O' levels



Publications

Journals:

Wedge, D., Ingram, D., McLean, D., Mingham, C. and Bandar, Z. (2006) On Global-Local Artificial Neural Networks for Function Approximation, IEEE Transactions on Neural Networks, 17(4), 942-952

Wedge, D., Ingram, D., McLean, D., Mingham, C. and Bandar, Z. (2005) On Neural Network Architectures and Wave Overtopping, Maritime Engineering, 158(MA3), 123-133

Proceedings:

Wedge, D., Hubbard, S., Gaskell, S.J., Kell, D.B., Lau, K.W. and Eyers, C. (2007) Peptide Detectability following ESI Mass Spectrometry: Prediction using Genetic Programming, GECCO 2007: Proceedings of the 9th annual conference on Genetic and Evolutionary Computation , In Press

Wedge, D., Ingram, D., McLean, D., Mingham, C. and Bandar, Z. (2005) A Global-Local Artificial Neural Network with Application to Wave Overtopping Prediction, Proceedings of the International Conference on Artificial Neural Networks (ICANN), Duch et al (eds.) 109-114


Miscellaneous




I have worked on a number of interesting projects over the years. These include the use of neural networks as a tool for lie detection through gesture analysis at Silent Talker and the development of web-based mapping software at Graticule. I also spent 7 years as a teacher of 5-7 year old children and 3 years as a 'house-husband', bringing up my oldest son.

I play chess regularly, captaining Hebden Bridge in the Calderdale League and representing Calderdale in the Yorkshire League. I am currently learning Chinese (Mandarin). I enjoy walking and running and completed the Pennine Way in 1987.



Downloads

The computer code used to run the Genetic Programming algorithm is written in the Java progamming language. Source code and class files are available for download here. Documentation in standard javadoc format is available here.

If you use this code in your research please credit me. Also any feedback (positive or negative) would be welcome.


Contact Details


E-mail:

Address: School of Chemistry
The University of Manchester
Manchester Interdisciplinary Biocentre
131 Princess Street
Manchester M1 7DN
United Kingdom

Telephone: +44 (0)161 306 5145

Personal | Research | Education | Publications | Miscellaneous | Downloads | Contact


Last update: 4 July 2007


Group





Back to the group's homepage.