To view this video please enable JavaScript, and consider upgrading to a web browser that supports HTML5 video


Machine Intelligence

Machine intelligence is closely related to artificial intelligence and machine learning. It deals with the issues and approaches related to developing automatic systems which exhibit intelligent behaviour and are able to act autonomously without human intervention. Often such behaviour can be inspired by biological systems, and technological developments aim to emulate this.

When creating machine systems with artificial intelligence, it is necessary to investigate the whole signal chain, from low-level sensing to high-level decision making. Sensors capture raw data which needs to be processed in order to provide meaningful information. Signal processing technologies are being developed which can detect patterns in acoustic signals and objects in visual video streams. Further processing through spatio-temporal modelling provides a reduction of the dataflow by efficient representation of information and knowledge, and logic and reasoning concepts can then lead to automated intelligent machine behaviour.

The applications of this technology are in robotics and intelligent human/computer interaction.


  • Robotics.
  • Machine vision.
  • Artificial intelligence.
  • Automated reasoning.
  • Signal processing.


Plus Icon Machine Intelligence Publications
  • Financial Time Series Prediction Using Spiking Neural Networks
  • Spiking Neural Networks for Financial Data Prediction
  • The Application of Ridge Polynomial Neural Network to Multi-Step Ahead Financial Time Series Prediction
  • Monekosso, D.N., Florez-Revuelta, F. and Remagnino, P. (2015) Guest Editorial Special Issue on Ambient-Assisted Living: Sensors, Methods, and Applications. IEEE TRANSACTIONS ON HUMAN-MACHINE SYSTEMS, 45 (5) October, pp. 545-549.
  • Bagheri Zadeh, P., Sheikh Akbari, A. and Buggy, T. (2015) DCT image codec using variance of sub-regions. DE GRUYTER OPEN: Open Computer Science, 5 (1) August, pp. 13-21. (Article)
  • Tawfik, H. and Anya, O. (2015) Evaluating practice-centered awareness in cross-boundary telehealth decision support systems. Telematics and Informatics, 32 (3) January, pp. 486-503. (Journal Article)
  • Elliott, J.R. (in press) Beyond an Anthropomorphic Template. Acta Astronautica.
  • Nefti-Meziani, S., Oussalah, M. and Soufian, M. (2015) On the use of inclusion structure in fuzzy clustering algorithm in case of Gaussian membership functions. Journal of Intelligent and Fuzzy Systems, 28 (4) January, pp. 1477-1493. (Journal Article)
  • Reid, D., Hussain, A.J. and Tawfik, H. (2014) Financial time series prediction using spiking neural networks. PLoS ONE, 9 (8) August. (Journal Article)
  • Brierley, C., Sawalha, M. and Atwell, E. (2014) Automatically-generated, phonemic Arabic-IPA Pronunciation Tiers for the Boundary-Annotated Qur'an Dataset for Machine Learning (version 2.0). LRE-Rel Workshop, Language Resources and Evaluation Conference (LREC 2014) 2014. (Conference proceedings)
  • Brierley, C., Sawalha, M. and Atwell, E. (2014) Tools for Arabic NLP: a case study in qalqalah prosody. Language Resources and Evaluation Conference (LREC 2014). (Conference proceedings)
  • Brierley, C., Sawalha, M., Atwell, E. and Dickins, J. (2014) Text Analytics and Transcription Technology for Quranic Arabic. International Conference on Islamic Applications in Computer Science and Technologies (IMAN). (Conference proceedings)
  • Mullen, R.J., Monekosso, D.N. and Remagnino, P. (2013) Ant algorithms for image feature extraction. Expert Systems with Applications, 40 (11) September,pp. 4315-4332. (Journal Article)
  • Monekosso, D.N. and Remagnino, P. (2013) Data reconciliation in a smart home sensor network. Expert Systems with Applications, 40 (8) June, pp. 3248-3255.(Journal Article)
  • Thida, M., Eng, H.L., Monekosso, D.N. and Remagnino, P. (2013) A particle swarm optimisation algorithm with interactive swarms for tracking multiple targets. Applied Soft Computing Journal, 13 (6) January, pp. 3106-3117. (Journal Article)
  • Kor, A. and Bennett, B. (2013) A Hybrid Reasoning Model for “Whole and Part” Cardinal Direction Relations. Advances in Artificial Intelligence.
  • Sheikh Akbari, A. and Bagheri Zadeh, P. (2012) Compressive sampling and wavelet-based multi-view image compression scheme. Electronics Letters, 48 (22) October, pp. 1403-1404. (Journal Article)
  • Thida, M., Eng, H.L., Monekosso, D.N. and Remagnino, P. (2012) Learning video manifolds for content analysis of crowded scenes. IPSJ Transactions on Computer Vision and Applications, 4 August, pp. 71-77. (Journal Article)
  • Grech, R., Monekosso, D.N. and Remagnino, P. (2012) Building visual memories of video streams. Electronics Letters, 48 (9) April, pp. 487-488. (Journal Article)
  • Hosseinian-Far, A., Pimenidis, E. and Jahankhani, H. (2012) Influence diagrams: Predictive approach in decision support systems. Intl Journal of Strategic Management and DSS, 17, pp. 16-22. (JOUR)
  • Brierley, C., Sawalha, M. and Atwell, E. (2012) Predicting Phrase Breaks in Classical and Modern Standard Arabic Text. Language Resources and Evaluation Conference (LREC 2012). (Conference proceedings)
  • Brierley, C., Sawalha, M. and Atwell, E. (2012) Open-Source Boundary-Annotated Corpus for Arabic Speech and Language Processing. Language Resources and Evaluation Conference (LREC 2012). (Conference proceedings)
  • Baxter, S. and Elliott, J. (2012) A SETI metapolicy: New directions towards comprehensive policies concerning the detection of extraterrestrial intelligence. Acta Astronautica, 78, pp. 31-36.
  • Akbari, A.S., Zadeh, P.B., Buggy, T. and Soraghan, J. (2012) Multiresolution, perceptual and vector quantization based video codec. Multimedia Tools and Applications, 58 (3) January, pp. 569-583. (Journal Article)

Case Study: Financial Time Series Prediction Using Spiking Neural Networks

The practical prediction accuracy of financial market behaviour attracts considerable interest from both research and business communities due to the potential for large financial gain associated with even very small efficiency gain margins.

A fraction of a percent improvement in prediction accuracy of stock or currency exchange data could result in different financial decisions being made and millions of pounds of financial gains or losses being avoided or significantly minimised. This is of particular interest to financial managers, investment advisors and auditors as it can support them in making critical financial decisions.

Despite the challenging nature of financial data forecasting, it remains a problem of high importance to solve since every improvement in the accuracy of financial data prediction over short, medium or long term can yield significant financial gains or improvements in financial risk management for businesses.

The need for more accurate and reliable prediction of financial market behaviour has motivated us to exploit more advanced types of neural networks.


Predicting future trends in the financial market such as currency exchange and stock market behaviour can be considered as a Financial Time Series Prediction task.

The prediction of financial time series is a complex scientific and engineering challenge as trends of the financial data can change significantly as a result of various correlated economic, political and even psychological factors. In addition, annual cycles such as the high demand for air travel in the summer months also affect exchange rates and fuel prices
Traditional methods for financial time series forecasting are based around standard statistical approaches and machine intelligence techniques such as support vector machine and fuzzy logic which achieve limited prediction capability due to the highly complex nature of most of the financial time series.

Neural network methods have demonstrated great potential due to their prediction ability, adaptability to different financial data prediction scenarios and robust prediction performance when dealing with complex, noisy and uncertain type of data or behaviour.

Multi-Layer Perceptrons (MLPs) and recurrent neural networks have been successfully applied to a broad class of financial market predictions and for some cases Higher Order Neural Networks also proved appropriate in terms of producing good prediction results.

Spiking Neural Networks (SNNs) have demonstrated the power in solving difficult problems in complex and dynamic environments and hence represented the main approach used in the project for financial time series prediction. We have applied a Spiking Neural Network (PSN) to solve non-stationary financial data prediction problems by exploiting the temporal characteristics of the spiking neural model.

Publications arising from research:

If you are interested in any of our services or have a question, please contact us

Back to Top Button