Research studentships in the School of Informatics

City University's School of Informatics undertakes world-class research in strategic areas that include artificial intelligence, geographic information systems, human-computer interaction, medical informatics, software and requirements engineering, and systems dependability.

The School is offering 11 full-time 3-year PhD studentships to support ongoing research on specific topics in these areas, to start in October 2004 and October 2005.

Application forms available from Alison Barrett at alisonb@soi.city.ac.uk. The closing date for applications is Friday 11th June at 17.00hrs. Each application must indicate the key area and topics (listed below) that the candidate is applying to, and describe their research background and interests in that area and topic - open-ended applications will be automatically rejected. Final award of the studentships will be made in early July. Candidates should be available for interview during the second half of June, and ensure that referees can be contacted during that period.

Candidates are expected to have at least a 2.1 Honours Degree or equivalent in a relevant subject. Conditional awards can be made if students have yet to receive their final degree classification. Each successful applicant will receive a bursary of £14,000 per annum for each year of the studentship. Studentships will start on the 1st October 2004 and, depending on availability, 1st October 2005. More details about the School of Informatics are available at www.soi.city.ac.uk

Note:there is a separate application process for non-studentship PhD study in the School of Informatics. Go to www.soi.city.ac.uk/research/phd/index.html to find out more.

ARTIFICIAL INTELLIGENCE

Effective Rule Extraction from Neural Networks

Much of the research on cancer nowadays involves the analysis of genes that might be expressed differently in cancer cells and normal cells. Understanding such complex differences between cancer and normal cells is key to the development of more effective cancer diagnosis techniques. In the case of breast cancer, the disease is identified in average eight years after its initial development stages. By identifying the crucial differences between normal and breast cancer cells, Bioinformatics could help reducing this gap between the initial stages of the disease and its diagnosis, what is fundamental to the successful treatment and indeed cure. Research is needed to investigate and develop algorithms for effective rule extraction from very large neural networks typically used to classify and distinguish normal cells from breast cancer cells. Supervisors include Dr Artur Garcez.

Animal-based Reinforcement Learning

Research is needed to develop new reinforcement learning algorithms based on recent advances in associative learning theories. Reinforcement learning techniques do still suffer from serious convergence and generalization problems. To solve such problems, research is needed to redefine reinforcement learning models so as to incorporate psychological bias that will guide exploration and exploitation in both single and multi-agent learning problems. Supervisors include Dr Eduardo Alonso.

A Formal Framework of Co-ordination in Open Multi-Agent Normative Systems

A number of logical models have been developed to represent various aspects of social interaction in open normative systems, namely, negotiation, argumentation, permissions, and rights. Research is needed to develop a comprehensive formal model of such multi-agent co-ordination techniques, using good background knowledge on formal logics for Artificial Intelligence, and, particularly, on the design and development of multi-agent systems. Supervisors include Dr Eduardo Alonso.

Principled Design for Nature-Inspired Optimisation

Novel nature inspired AI optimisation methods such as ant systems, particle swarm optimisation and artificial immune systems are increasingly finding application but their design is somewhat ad hoc. Research at City has looked at how to design optimisers in a principled way, based on the domain knowledge that they exploit. This has already been applied to genetic algorithms, but needs extending to other nature-inspired methods.

A student with confident mathematical skills is needed as the design framework makes heavy use of discrete maths and first-order logic. Supervisors include Dr Andrew Tuson.

GEOGRAPHIC INFORMATION SYSTEMS

Interactive Maps for Information Access

Highly interactive maps are being used increasingly by academic researchers in the geosciences to explore complex multivariate spatio-temporal data sets. This process of geo-visualization prompts thought and plays a key role in the advancement of science and construction of knowledge. The techniques can also be used as a metaphor for information access, providing a wide range of users with an exploratory interface to known information using map-based techniques. But how successful is this approach to information provision? Which techniques are appropriate, and under what circumstances? To what extent can we use interactive maps to provide spatial and aspatial information to a wide audience through a visualization interface? Interdisciplinary research will be undertaken to address these and related questions by drawing upon local expertise in cartography, geovisualization, Internet technologies, information provision for consumers, interaction design and usability evaluation in the context of impacts of the Internet upon industry, government and society. Supervisors will include Dr Jason Dykes.

Using Geographic Information to Model Ethnophysiographic Representations of Landscapes

Traditional analysis of digital elevation models (DEMs) is suitable for partitioning surfaces into discrete landscape features such as channels, peaks, pits and passes. However features identified in this way often show artefacts of the data, scale and GIS used to extract them, and frequently do not correspond with our own cognitive models of the landscape. This PhD research topic will use some of the recent ideas to emerge from the field of 'Ethnophysiography' combined with novel GIS-based surface processing techniques to produce a more human-centred approach to modelling landscapes and their characteristics. Proposed research will entail developing and implementing new DEM processing operations as well as empirical comparison with named Landscape feature and user-perceived models of terrain. Supervisors will include Dr Joseph Wood.

Representing and Analysing Mobile Phone Trajectories in 4D Space-Time

A variety of 'location-enabled' mobile devices that can record their trajectories through space and time are now on the market, ranging from handheld GPS, through mobile phones to vehicle tracking systems. If collected, this stream of data needs to be data based, summarised, simplified and analysed to preserve privacy, reduce data volumes and extract useful location trends or 'significant' points. 'Error bars' based on GPS quality of service indicators or building density are also to evaluate the stream of data. The proposed research will develop algorithms to compress, characterise, generalise and evaluate the data so that the underlying navigation, tracking, guiding or gaming applications can use the data stream effectively. Supervisors will include Professor Jonathan Raper.

HUMAN-COMPUTER INTERACTION

Designing Interactive Systems for the Elderly and Disabled

It is estimated that there are more than 6.5 million disabled adults in Britain and many more millions world-wide have physical, sensory or cognitive limitations that make interacting with computers difficult. The number of people with disabilities is expected to increase significantly in the next decade as the UK's and the world's populations are rapidly growing older, and the number of computer users of old age also increases exponentially. PhD studentships are available to continue existing world-class research at City to investigate human-computer interaction issues for people with disabilities and the elderly. One aspect of particular interest is to investigate the accessibility of the WWW for disabled users and to develop and evaluate an accessible development lifecycle for websites. Supervisors include Professor Helen Petrie and Dr Panayiotis Zaphiris.

E-Learning and Usability

As e-learning provision continues to expand in all education sectors, usability has emerged as a critical factor in the success or otherwise of e-learning systems. The usability of these systems has a significant impact on the students’ learning experience and the extent to which they can achieve their learning outcomes. City is engaged in research into navigation strategies for effective e-learning, information visualisation techniques for digital libraries and innovative design and evaluation techniques for elearning systems. PhD studentships are available to continue and develop this leading-edge research. Supervisors include Stephanie Wilson and Dr. Panayiotis Zaphiris.

Safety-Critical Socio-Technical Systems

City University is working with London hospitals on research aimed at reducing the recurrence of adverse events in the NHS and improving safety for patients. We are developing a holistic method for modelling clinical environments and using this to design information appliances and resources using a distributed cognition approach. We are also interested in the process of shift changeovers, both in the medical arena as well as other safety-critical industries. PhD studentships are available to work on the modelling of, and designing for, clinical and other safety-critical environments. Studentships are also available to continue investigations into the role of technology in facilitating shift changeovers. Supervisors include Dr. Julia Galliers and Stephanie Wilson.

Virtual Ethnography of Computer Mediated Communication

The study of Computer Mediated Communication (CMC) synthesises skills from a multitude of disciplines (e.g. ethnography, sociology, computing). Virtual ethnography (the study of online communication through the use of ethnographic techniques) has received a lot of attention and when combined with analytic techniques (e.g. social network analysis) can be a valuable synthesis of methodologies for evaluating and analysing online communities. PhD studentships are available to continue this research, especially in investigating the use of Social Network Analysis for analysing online communities. Supervisors include Dr Panayiotis Zaphiris.

Usability Evaluation

Usability evaluation techniques are a central component of user-centred design methods and make a significant contribution to the usability of interactive systems. An on-going programme of research at City has been concerned with investigating the effectiveness of existing usability evaluation techniques and with developing new approaches. New paradigms, such as ubiquitous and mobile computing, e-learning environments and systems for elderly and disabled users, pose new challenges for usability evaluation that cannot be met by current evaluation techniques. PhD studentships are available to continue this research, with a particular emphasis on investigating the important usability issues for these new interaction paradigms from a theoretical perspective and on developing practical evaluation strategies that will meet their challenges. Supervisors include Stephanie Wilson and Professor Helen Petrie.

Navigation

Navigation in electronic information spaces is often reported as problematic, for example users are frequently unable to determine where they are, how they got there and where they can go to next. PhD studentships are available to research a range of navigation issues, including developing theoretical models of navigation, understanding and developing appropriate user support for navigation on the WWW and investigating navigation in e-learning environments. Supervisors include Professor Helen Petrie, Stephanie Wilson and Panayiotis Zaphiris.

MEDICAL INFORMATICS & DIGITAL INFORMATION IN HEALTHCARE

Studentships are available for 4 topics in medical informatics.

WINORE - Wearable computers IN the Operating Room Environment

This research will develop an enhanced patient monitoring system, based on wearable computer technology, for use in the operating theatre. The modern operating theatre is a busy environment with many people and the ever-increasing presence of high technology. One solution could be to use wearable computer technology to provide the clinicians with a portable system that provides easy to view data at the same time as looking at the patient. The work will involve researching, building and configuring a wearable computer system, researching and implementing the interaction aspects of providing clinical data in an operating theatre and evaluating the system, ultimately in a clinical environment. Supervisors will include Dr Peter Weller and Professor Helen Petrie, will clinical partners from the Royal London Hospital.

MADGIC: Modelling of Anastomatic Device Grafts In Cardiac surgery

This research will develop a prediction tool to enable cardiac surgeons to determine the performance and lifetime of coronary bypass surgery. With the advancement of minimally invasive methods and robotic cardiac surgery, suture less anastomotic devices have been recently introduced into surgical practice, generating the need for a support tool to assist the cardiac surgeon to determine the nature, expected performance and life span of a graft for each patient. The work will involve researching and modelling the blood flow in a bypass graft, using angiograms and measured data to verify the model, developing the final support tool, and evaluating the system, ultimately in a clinical environment. Supervisors will include Dr Peter Weller and Professor Ewart Carson, with clinical partners from the London Chest Hospital.

Digital Information Use by Health Professionals

Healthcare professionals increasingly use digital information, especially in regard to the use of NeLH (National Electronic Library for Health). Recent research shows that medical professionals might be taking on some of the characteristics of the health consumer in the way that they search for information – increased choice and search engines are beginning to change time-honoured information practices. A paradigm shift in behaviour might be occurring, have little data is available to validate this claim, which needs testing using a triangulated methodological mix, employing logs and online questionnaires. A PhD studentship is available to investigate this research issue. Supervisors will include Professor David Nicholas and Dr Ian Rowlands

Internet Digital Libraries

Internet digital libraries provide services to different user groups. A one-size-fits-all solution cannot meet the needs of all users and user groups. Research is needed to investigate the issues of user customisation of information delivery in digital libraries bridging the research areas of agent-based technologies, digital libraries evaluation, the semantic web, push and pull information dissemination strategies, and knowledge management. Supervisors include Dr Abdul Roudsari.

SOFTWARE AND REQUIREMENTS ENGINEERING

Studentships are available for 6 topics in which City’s software and requirements engineering research is world-class.

Service-Centric Systems Engineering

The development of software systems as dynamic compositions of web services is emerging as a powerful new software engineering paradigm. City University has strong interests and on-going research activities in this area – in particular a EU-funded Framework VI Integrated Project – which are related to web-service discovery, management of conflicts in service composition, and run-time service monitoring for quality assurance. PhD studentships are available for research related to the above themes. Supervisors include Dr. George Spanoudakis, Dr. Andrea Zisman and Professor Neil Maiden.

Scenario-Driven Software Engineering

Scenarios, stories and use cases are one of the most influential techniques to have emerged in software development over the last decade. City has been at the leading edge of this research, developing innovative web-enabled scenario generation and walkthrough tools such as ART-SCENE (www.soi.city.ac.uk/artscene) and evaluating their use in large-scale projects in air traffic control. PhD studentships are available to continue this research, in particular to research how multi-media scenarios can inform requirements discovery and modelling and how agent-based scenario animations enhance scenario walkthroughs. Supervisors include Professor Neil Maiden.

Creative Requirements and Software Design

Specifying and design software-based products, from mobile phones to air traffic control systems, is increasingly a creative process. Innovative products are essential to establish new markets and advance technologies. However, most development methodologies support elicitation, analysis and management rather than innovation. City University has been working with the Atlantic Systems Guild to establish and evaluate methods and tools for specifying and designing more innovative systems, based on theories of creativity from other disciplines. Results have been successfully applied in air traffic control projects. PhD studentships are available to continue this work, in particular to innovate new scenario-driven tools that encourage creative thinking for more innovative systems design. Supervisors include Professor Neil Maiden.

Traceability

Software Traceability has been recognised as an important activity in software system development and a factor of software quality. City University has significant research activities in this area related to the study of different types of traceability relations and development of techniques and tools for the representation, generation, and maintenance of these relations. Of particular interest is the study of traceability in the context of software product families, the automatic generation of traceability relations and the use of machine learning techniques for this purpose, and the establishment of traceability for analysis related to security and safety requirements. PhD studentships are available to continue research in these areas. Supervisors include Dr Andrea Zisman and Dr George Spanoudakis.

Inconsistency Management

The management of inconsistencies between software models is a complex process involving different activities including the detection of overlaps and the discovery, handling and tracking of inconsistency between such artefacts. City University has an excellent track record of research in this area and has produced methods and tool support for the management of inconsistencies including the Reconciliation+ method of referential system modelling and goal modelling approaches. PhD studentships are available to continue research in this area. Topics of research include the refinement of inconsistency management techniques for validating web service systems and for aligning the business processes underpinning such systems. Supervisors include Dr Andrea Zisman and Dr George Spanoudakis.

Ubiquitous Requirements Engineering

Establishing the requirements for future systems is work that must increasingly be done in the work place, to capture the context of work and the environment in which the system must operate. However, most requirements engineering tools remain desktop-based, unable to support this new work. In contrast mobile computing technologies such as PDAs provide new platforms and aids for requirements tasks, from acquiring requirements in context to participating in negotiation meetings remotely. PhD studentships are available to explore how ubiquitous computing can enhance future requirements processes, and develop new theories and models of these requirements processes. Supervisors include Professor Neil Maiden.

SYSTEMS DEPENDABILITY

Dependable Services with Diverse Off-The-Shelf Software

Software fault-tolerance has been known for 30 years but the take up by industry has been limited so far due to various factors, not least because of the high cost of bespoke development of multiple software channels. With the success of standardisation and as a result the market for off-the-shelf software this major impediment has been overcome - multiple channels with similar functionality are available as relatively cheap commodities for a very large number of services. In a recently completed EPSRC project we have started an effort of practical utilisation of diversity with complex off-the-shelf software such as SQL servers and have reported empirical evidence that diversity can improve both their dependability and performance. Follow-on research is needed to implement a middleware for database replication to work with two diverse open-source servers and by expanding in the emerging area of Web-services. For dependable services we intend to adapt the ideas of the probabilistic dependability assessment that City has developed over the years, mainly for safety-critical industries. Supervisors include Dr Peter Popov.

Fault Tolerance and Diversity in Human-Computer Systems

Assessing and improving the dependability of computer systems that support human activities requires studying the wider systems formed by users and computers. Important aspects of such systems are the fault tolerance provided by humans and machines for each other's errors, and the changes that computer support causes in human performance. Previous research applied an interdisciplinary approach involving psychology and reliability engineering and both descriptive and statistical methods. Case study results on a support system for cancer screening detected useful new facts to support system design and assessment. More research is to continue this work with extensions to other case studies and to develop generalisations of the methods applied. Supervisors include Professor Lorenzo Strigini.

Safety Cases, Confidence, and Diversity

When computer-based systems provide critical services (e.g. when human life depends upon them, or their failure can result in serious social disruption), it is important to know that they are sufficiently dependable: that confidence in them is not misplaced. New research is needed to investigate how such confidence can be obtained from the disparate evidence that is usually available, including the use of 'diverse' arguments that tolerate human error and probabilistic modeling. Supervisors include Professor Bev Littlewood and Professor Robin Bloomfield.

Dependability Modeling for Security

City has successfully developed models and methods for assessing the safety and reliability of software systems. This probabilistic approach facilitates quantitative assessment, and thus allows rigorous risk assessment of complex socio-technical systems. We are now extending this work to take account of security failures. Probabilistic approaches to security assessment are new, and City is in the forefront of work attracting considerable interest as a result of recent international concerns. Supervisors include Professor Bev Littlewood and Professor Robin Bloomfield.