School of Computing and Mathematics
Explore this Section
- Faculty of Natural Sciences /
- School of Computing and Mathematics /
- Research /
- Research Groups /
- Computational Intelligence and Cognitive
Computational Intelligence and Cognitive Science
The group’s research focuses on simulating and applying biologically inspired and conventional machine learning models of intelligent systems, in order to widen our understanding of their fundamental properties and, through our interdisciplinary collaborations with staff within the Research Institute for the Environment, Physical Sciences and Applied Mathematics (EPSAM), in the Institute for Science & Technology in Medicine (ISTM), at other UK and international universities, and industry, to extend their range of applications.
Dr Aston, Mr Borg, Dr Butcher, Dr Channon, Dr Day, Dr Kyriacou and Mr Stanton use computational models and simulation to address fundamental issues surrounding evolution, machine intelligence, robotics and the emergence of complexity, as well as helping to solve real-world problems. Mr Collins, Dr Fletcher, Dr Lam and Mr Smith’s research is directed towards automated object recognition, focussing particularly on medical applications, working directly with biomedical science and bioengineering colleagues in ISTM, and on analytical techniques (including machine learning) to enable classification and visualisation of large datasets. This research has high impact across a range of industrial and medical applications.
Our EPSRC-funded research into evolutionary dynamics has led to the discovery that, as population size falls, the 'critical mutation rate' above which fitter alleles are lost transitions unexpectedly from near-constant (the previous assumption in evolutionary biology) to drop exponentially for small populations, leaving them spiralling toward extinction (Channon et al. 2011; Aston et al. 2013). Current research is using on the implications this new understanding has for species under threat of extinction, and its potential to aid population management and prevent extinctions. With regard to the evolution of neurally controlled agents, we have identified a significant weakness in the standard approaches taken in neuroevolutionary bio-inspired robotics, and we presented a solution that dramatically improves the quality of incremental evolutionary search for complex behaviours (Stanton & Channon 2013). EPSRC funded work on the evolution of DNA sequences to bind to transcription factors and other proteins has led to a new understanding of optimal mutation rate control (in theory, simulation and biology: Belavkin et al. 2011; Krasovec et al. 2014, Nature Communications). Following this, the project team has been awarded a BBSRC grant to advance knowledge of how bacteria evolve antibiotic resistance, with a view to helping to develop ways of preventing the spread of antibiotic resistant 'superbugs'. The group also undertakes research into the evolution of social learning, with two significant successes: we have presented the first example of a behaviour inaccessible to incremental genetic evolution alone being evolved through the addition of cultural transmission (Borg et al.2011); and we have provided the first definitive answer to the previously open question of whether or not the “variability selection hypothesis” is sufficient to explain the adoption of social learning in increasingly variable environments (Borg & Channon 2012).
Dr Day’s EPSRC-funded research includes the use of neural networks for the automated element-specific detection of sensitive substances illuminated by the latest X-ray scanning equipment (Day 2009). This work is highly collaborative with EPSAM colleagues in Chemical Sciences (Haycock, Austin) and Criminology (Kearon) and is the first to show that self-organising neural networks can effectively be used to make chemical-element determinations based on a target’s transmitted X-ray signature. This has applications in domains as diverse as airport security scanning, provenance of ancient artefacts, and tracing illegally modified firearms – a socially important application that led to the UK Forensic Science Service supporting this work. A separate international collaboration, partially funded by SciSite Ltd. (now Scicorr Ltd.), has resulted in the development of a new reservoir computing neural network (Butcher, Verstaeten, Schrauwen, Day & Haycock 2013) that is better able to diagnose the structural integrity of reinforced concrete structures based on non-invasive electromagnetic survey data (Butcher, Day, Haycock, Verstraeten & Schrauwen 2013). Work is continuing to develop this very promising strand of research, potentially extending the work to employ multi-sensor data fusion techniques to supplement the analysis of electromagnetic-survey data. Other collaborative research with colleagues in Analytical Chemistry and Forensic Science (Drijfhout, Adam) is developing neural network tools and techniques to assist law enforcement and forensic investigators in their attempts to determine more accurately and more readily the post-mortem interval in homicide cases using forensic entomology data. The team has demonstrated for the first time that a self-organising map can be systematically and accurately used to automate the clustering and classification of the profiles of insect larvae varying in age from one to nine days old, without the need for subjective expert interpretation of the data (Day 2013).
Dr Fletcher's research topic is syntactic pattern recognition. The aim is to develop a method for robust recognition of complex recursively structured geometric patterns, in the presence of noise, vagueness, occlusion, distortion and overlapping of patterns.
Dr Kyriacou's biologically inspired robotics research focuses on biological mechanisms of navigation (Kyriacou 2011, Kyriacou et al. 2011, Kyriacou 2012). His other robotics research is concerned with non-biological mechanisms of robot navigation (Kyriacou, Styles & Toon 2008); human-robot interaction, computer vision, analytical robotics and systems modelling (Kyriacou, Nehmzow, Iglesias & Billings 2008); and novel methods for teaching computer programming using robots (Major 2011).
Medical applications of Dr Lam’s BBSRC-funded work on automated object recognition benefit greatly from productive collaborations with biomedical scientists and bioengineers in ISTM (Richardson, Yang, Sulé-Suso), and include the identification of tissue boundaries in computer tomographic (CT) scans, new fractal algorithms to characterise the quality of transplanted cell growth from post-operative biopsies, and algorithms to improve not just the tracking of cells but also the auto-focus method used in high throughput phase contrast microscopy. One project (Lam, Collins and Smith) developed new (multi-scale) fractal algorithms to model and characterise the quality of transplanted cell growth from post-operative biopsies (Smith 2013) in order to facilitate a high quality medical capability for large-scale, in-vitro patient-specific generation of cartilage growth for the treatment of arthritis. This work has attracted the interest of oncologists and has led to a joint publication in Laboratory Investigation (Nature Publishing Group) with colleagues in ISTM. Research on the classification and visualisation of ‘big data’ has led to the development of the first practical exploratory framework that unified Independent Component Analysis and Projection Pursuit (Lam & Emery 2009; Smith 2013). Research funded by the company Forensic Pathways on source camera identification in digital image forensics has led to the development of a benchmarking signal/component separation procedure based on the eigen-analysis approach to non-parametric spectral decomposition, which was shown to better characterise and classify sensor pattern noise in low SNR conditions typical of most mobile/phone devices (Lam 2011).
At Keele, the importance of research-led and research-informed teaching is also recognised. Computational Intelligence is a strong thread within our undergraduate programmes, with Computational Intelligence modules embedded within our single honours Computer Science programme and a specialist dual honours programme in Smart Systems.
Professor Peter Andras (email@example.com)
Within Keele, the group has developed productive collaborations with colleagues in Analytical Chemistry, Forensic Science, Physics and Applied Geophysics in EPSAM, with colleagues in biomedical sciences, bioengineering, neuroscience and physiotherapy in ISTM, and with Kearon in Criminology. Drs Day and Kyriacou continue to collaborate with colleagues from Psychology, Neuroscience and Physiotherapy to explore how robotics can be used to assist subjects with locomotive impairments (e.g. following a stroke) or neurodegenerative disorders such as Parkinson’s Disease. Dr Day’s productive collaborations with Haycock and Austin (Chemical Sciences) and Kearon (Criminology) and Drijfhout and Adam (Analytical Chemistry and Forensic Science) are described above.
Drs Day and Butcher (Life Sciences) have established a productive collaboration (two journal papers, 2012 and 2013) with colleagues from the University of Ghent in Belgium. The Ghent group are one of the internationally pre-eminent groups in the development and application of reservoir-computing neural-network techniques in engineering and robotics.
Dr Kyriacou collaborates with colleagues from Keele's School of Health and Rehabilitation on data mining and modelling in the area of stroke rehabilitation; modelling human gait for the purposes of informing interventions to help patients with disabilities related to their walking; and modelling upper arm movement and function in order to inform the design of prosthetic limbs. Dr Kyriacou is also investigating the data mining of NHS primary care data in order to detect charging anomalies/errors caused when NHS services charge the local NHS authority, and collaborating with colleagues from ISTM and life sciences on data mining and modelling in the area of stem cell research, particularly in relation to in-vitro cultures of neuronal stem cells.
Through an EPSRC-funded project, Dr Channon worked (2010-2013) with partners Drs Chris Knight, Rok Krasovec, Roman Belavkin and John Aston, from the Universities of Manchester, Middlesex and Warwick (now Cambridge), to advance understanding of (amongst other things) the evolution of DNA sequences and their bindings to transcription factors and other proteins. This ongoing collaborative project involved both the evolution of wetware (biological) DNA and the much faster evolution of sequences in computer experiments using tables of DNA-to-protein binding affinities (Belavkin et al. 2011) and Nvidia's CUDA many-core (GPGPU) technology. The team began a follow-on BBSRC-funded project on mutation rate plasticity in Febrauary 2014.
Dr Lam and Mr Collins have strong research collaborations with medical researchers and practitioners, and with forensic scientists both at Keele and in industry. Active collaborators in ISTM include Prof. James Richardson and his research team at the Robert Jones and Agnes Hunt Orthopaedic Hospital in Oswestry, with whom research is focused on automated object recognition (Smith 2013), and Drs Ying Yang and Josep Sulé-Suso focusing on non-destructive analysis of cells. Application-driven research in the forensic science domain has been in collaboration with Criminal Records Direct Ltd. (formerly Assuramed Ltd.) and with Forensic Pathways Ltd., as well as with forensic science researchers at Keele University (Lam et al. 2011).
Krasovec R, Belavkin RV, Aston JAD, Channon A, Aston E, Rash BM, Kadirvel M, Forbes S, Knight CG. 2014. Mutation rate plasticity in rifampicin resistance depends on Escherichia coli cell-cell interactions. Nature Communications, vol. 5, article 3742. doi
Aston E, Channon A, Day C, Knight CG. 2013. Critical Mutation Rate Has an Exponential Dependence on Population Size in Haploid and Diploid Populations. PLOS ONE, vol. 8(12), Article ARTN e83438. doi full text
Butcher, J, Verstaeten, D, Schrauwen, B, Day CR, Haycock, P. 2013. Reservoir computing and extreme learning machines for non-linear time-series data analysis. Neural Networks. doi
Butcher, JB, Day CR, Haycock, PW, Verstraeten, D, Schrauwen, B. 2013. Defect Detection in Reinforced Concrete Using Random Neural Architectures. Computer-Aided Civil and Infrastructure Engineering. doi
Day CR, Butcher, JB, Moore, HE, Drijfhout, FP, Adam,CD. 2013. Artificial Neural Network analysis of hydrocarbon profiles for the ageing of Lucilia sericata for Post Mortem Interval estimation. Forensic Science International, vol. 232(1-3), 25-31. doi
Smith, WA, Lam K, Collins, D, Tarvainen, J. 2013. Estimation of Depth Map using Image Focus: A Scale-Space Approach for Shape Recovery. doi
Stanton A and Channon A. 2013. Heterogeneous complexification strategies robustly outperform homogeneous strategies for incremental evolution. In P. Lio, M. Orazio, G. Nicosia, S. Nolfi & M. Pavone (Eds.). Advances in Artificial Life (pp. 973-980). MIT Press. doi
Kyriacou T. 2012. Using an Evolutionary Algorithm to Determine the Parameters of a Biologically Inspired Model of Head Direction Cells. Journal of Computational Neuroscience. doi
Li J, Li B, Wo T, Hu C, Huai J, Liu L, Lam KP. 2012. CyberGuarder: A virtualization security assurance architecture for green cloud computing. Future Generation Computer Systems – The International Journal of Grid Computing and eScience, vol. 28(2), 379-390. doi
Quintía P, Iglesias R, Regueiro CV, Rodríguez M, Kyriacou T. 2012. Selecting the most relevant sensors in a wall following behavior. Workshop of Physical Agents 2012 (WAF 2012). Santiago de Compostela, Spain. link
Belavkin RV, Channon A, Aston E, Aston J, Knight CG. 2011. Theory and practice of optimal mutation rate control in hamming spaces of DNA sequences. In T. Lenaerts, M. Giacobini, H. Bersini, P. Bourgine, M. Dorigo & R. Doursat (Eds.). Advances in Artificial Life, ECAL 2011: Proceedings of the Eleventh European Conference on the Synthesis and Simulation of Living Systems (pp. 85-92). Heidelberg: MIT Press. doi full text
Borg JM, Channon A, Day C. 2011. Discovering and maintaining behaviours inaccessible to incremental genetic evolution through transcription errors and cultural transmission. In T. Lenaerts, M. Giacobini, H. Bersini, P. Bourgine, M. Dorigo & R. Doursat (Eds.). Advances in Artificial Life, ECAL 2011: Proceedings of the Eleventh European Conference on the Synthesis and Simulation of Living Systems (pp. 101-108). Heidelberg: MIT Press. doi full text
Channon A, Aston E, Day C, Belavkin RV, Knight CG. 2011. Critical mutation rate has an exponential dependence on population size. In T. Lenaerts, M. Giacobini, H. Bersini, P. Bourgine, M. Dorigo & R. Doursat (Eds.). Advances in Artificial Life, ECAL 2011: Proceedings of the Eleventh European Conference on the Synthesis and Simulation of Living Systems (pp. 117-124). Heidelberg: MIT Press. doi full text
Kyriacou T, Butcher J, Day C. 2011. A Model of Head Direction Cells with Changing Preferred Head Direction. 4th International Workshop on Evolutionary and Reinforcement Learning for Autonomous Robot Systems (ERLARS 2011). Berlin, Germany. link
Kyriacou T. 2011. An Implementation of a Biologically Inspired Model of Head Direction Cells on a Robot. Towards Autonomous RObotic Systems (TAROS. doi
Lam K, Soobhany, AR, Leary, R. 2011. On the Performance of Li’s Unsupervised Image Classifier and the Optimal Cropping Position of Images for Forensic Investigations. International Journal of Digital Crime and Forensics, vol. 1(3). doi
Major L, Kyriacou T, Brereton P. 2011. Experiences of Prospective High School Teachers Using a Programming Teaching Tool. 11th Koli Calling International Conference on Computing Education Research (Koli Calling ’11). Finland. doi
Major L, Kyriacou T, Brereton OP. 2011. Systematic Literature Review: Teaching Novices Programming Using Robots. Evaluation and Assessment in Software Engineering (EASE 2011). Durham, UK. doi
Major L, Kyriacou T, Brereton P. 2011. Simulated Robotic Agents As Tools To Teach Introductory Programming. International Technology, Education and Development Conference (INTED 2011) (pp. 3837-3846). Valencia, Spain. ISBN 978-84-614-7423-3. link
Nasirudeen AMA, Wong HH, Thien P, Xu S, Lam K-P, Liu DX. 2011. RIG-I, MDA5 and TLR3 Synergistically Play an Important Role in Restriction of Dengue Virus Infection. PLOS NEGLECTED TROPICAL DISEASES, vol. 5(1), Article ARTN e926. doi
Austin JC, Day CR, Kearon AT, Evans DL, Haycock PW. 2010. Comparison method to differentiate between painted objects using polychromatic X-rays. INSIGHT, vol. 52(3), 140-143. doi
Butcher, J, Verstraeten, D, Schrauwen, B, Day CR, Haycock, P. 2010. Extending reservoir computing with random static projections: a hybrid between extreme learning and RC. link
Butcher, JB, Verstraeten, D, Schrauwen, B, Day CR, Haycock, PW. 2010. Pruning reservoirs with random static projections. doi
Pijanka J, Sockalingum GD, Kohler A, Yang Y, Draux F, Parkes G, Lam KP, Collins D, Dumas P, Sandt C, van Pittius DG, Douce G, Manfait M, Untereiner V, Sulé-Suso J. 2010. Synchrotron-based FTIR spectra of stained single cells. Towards a clinical application in pathology. Lab Invest, vol. 90(5), 797-807. link doi
Lam K and Smith, WA. 2010. Exploratory Analysis of UV-Vis Absorption Spectra. doi
Austin JC, Day CR, Kearon AT, Haycock PW. 2009. Single element mapping in radiography. X-RAY SPECTROMETRY, vol. 38(6), 492-504. doi
Butcher JB, Lion M, Day CR, Haycock PW, Hocking MJ. 2009. A Low Frequency Electromagnetic Probe for Detection of Corrosion in Steel-Reinforced Concrete. In M. Grantham & C. Majorana (Eds.). Concrete Solutions (pp. 446-451). CRC Press. doi
Day CR, Austin JC, Butcher JB, Haycock PW, Kearon AT. 2009. Element-specific determination of X-ray transmission signatures using neural networks. NDT & E International, vol. 42(5), 446-451. doi
Lam KP and Fletcher P. 2009. Concurrent grammar inference machines for 2-D pattern recognition: a comparison with the level set approach. In JT. Astola, KO. Egiazarian, NM. Nasrabadi & SA. Rizvi (Eds.). Image Processing: Algorithms and Systems VII (SPIE vol. 7245). SPIE-IS&T. doi
Lam KP and Emery R. 2009. Image Pixel Guided Tours: A Software Platform for Non-Destructive X-ray Imaging. Image Processing: Algorithms and Systems VII (SPIE vol. 7245). doi
Austin JC, Day CR, Kearon AT, Valussi S, Haycock PW. 2008. Characterisation of metallic powder impregnated pastes using polychromatic X-radiography. INSIGHT, vol. 50(10), 550-553. doi
Kyriacou T, Nehmzow U, Iglesias R, Billings SA. 2008. Accurate Robot Simulation Through System Identification. Robotics and Autonomous Systems (vol. 56, p. 1082–1093). doi
Kyriacou T, Styles P, Toon S. 2008. Robot Localization Using Seismic Signals. TAROS 2008 Towards Autonomous Robotic Systems. Edinburgh, United Kingdom. link