Computational Intelligence Group
The RPLOT program provides an informative way of analysing 15N relaxation data
This group brings together interdisciplinary researchers investigating the interface between computer science and the domains of bioscience and cognition. In terms of applying computation to other domains, we have experts in investigating the modelling of gene expression and modelling of human attention, emotions and reasoning. From the perspective of applying biological metaphors to computation, we research new computational methods such as genetic algorithms and swarm intelligence.
The group also develops novel techniques for data mining, visualization and simulation. These use the results of interdisciplinary research to find solutions for computationally expensive problems. The methods we develop are also applied to the bioscience and cognitive domains to model systems and to explore the data within them.
Areas of Research Activity
Members are engaged in the following areas of research:
|*||Application of computational simulations in biology and medicine.||*||Bio-inspired computing - genetic algorithms, swarm intelligence, artificial neural networks.|
|*||Systems biology and bioinformatics||*||Theory and applications of information visualisation|
|*||Data mining and knowledge discovery||*||Salience sensitive control|
|*||Attention, affect and addiction||*||Probabilistic planning and reinforcement learning|
Some suggestions of possible postgraduate research projects are also available.
The Computational Intelligence Group’s research can be broken down into modelling, novel computing paradigms and data analysis.
- Modelling: David Barnes and Dominique Chu are experts in applying computation to investigating the modelling of gene expression, and work with bioscientists to develop computational models of biological systems behaviour. Howard Bowman, Colin Johnson and Caroline Li have examined models of human attention, creativity, emotions and reasoning. This research has focussed on modelling cognition supported by EEG data provided by Psychologists, and Monte Carlo methods have revealed new insights for detecting neural markers of conscious activity.
- Novel Computing Paradigms: Colin Johnson and Fernando Otero apply biological metaphors to computation, and research new computational methods such as genetic algorithms and swarm intelligence in areas including data analysis, music technology and bioinformatics.
- Machine Learning and Data Analysis: Alex Freitas and Fernando Otero have developed novel techniques for the classification task of machine learning or data mining, including new techniques for hierarchical classification, multi-label classification, and bio-inspired algorithms for classification. These are often applied to biological data, including data on the biology of ageing. Marek Grzes has developed probabilistic methods for planning (e.g. based on Markov decision processes) and reinforcement learning as well as worked on applications of those algorithms in the area of assistive technology (i.e. intelligent systems for people with disabilities) and natural language processing (especially chatbots).Peter Rodgers has developed new algorithms for information visualization to explore data within complex systems, and these have been applied in areas such as business planning, transport, bioscience and medicine.
- Computational Creativity: Colin Johnson and Anna Jordanous have worked on computational creativity, that is the use of computational intelligence techniques to produced results which an outside observer would deem to be creative. This work has involved both the development of creative systems in the area of music, the development of means to evaluate creativity, and the development of theories to understand creativity in complex human-machine systems.
The Computational Intelligence Group will continue to emphasise cross-disciplinary, collaborative work. Our research plans in the forthcoming period include:
- In the area of systems biology, we will continue work to understand the thermodynamic limits of computing in living cells. We will analyse data from systems pharmacology and the systems biology of ageing using new probabilistic data mining techniques.
- Neurological modelling work will use EEG data to study the relationship between conscious awareness and attention, and will be applied to forensics, such as lie detection.
- Our visualization efforts will examine the communication of risk, as well as hybrid text and visualization methods for categorical data.
- In the area of probabilistic planning, we will focus on the development of algorithms for decision-theoretic planning and machine learning with applications to intelligent tutoring systems, natural language processing, assistive technologies, and games.