Dear all,

Please see below for information about an opportunity to study for a PhD on the interface between evolutionary computation human computer interaction at the University of Exeter, UK. We would greatly appreciate it if you could share it with anyone who may be interested.

"Human Computer Interaction for Evolutionary Computation"

Evolutionary algorithms (EAs) are nature-inspired techniques that are widely used to optimise complex problems in science and industry. Interactive EAs incorporate an expert from the problem domain to include their knowledge into the optimisation process, which can greatly improve results; a problem, however, is that EAs are highly complex systems and do not work in terms with which a typical user is familiar, making them difficult to relate to. As such, developing intuitive systems for human-algorithm interaction is vital.  This work will use the principles of human-computer interaction (HCI) to develop new mechanisms of interaction between the user and algorithm, including the capturing of domain expertise and visualisation of solution sets.  It will also seek to enable interaction with the whole evolutionary processes, particularly the mechanisms used to generate new solutions. Examples of areas that will be investigated include:

 * Interaction with the application of crossover and mutation operators, and the effect their use has on the population.
 * Visualisation within dynamic and uncertain environments – illustrating, for example, how the fitness landscape changes over time.
 * The relationship between solutions and their corresponding objective(s).
 * Application to other nature-inspired techniques, and related algorithms such as hyper-heuristics.

Evaluation of the human-algorithm interaction methods will be conducted rigorously using HCI research methods drawn from an interdisciplinary supervision team from Computer Science and Psychology.  Since an objective of the project is to make EAs more intuitive, a natural consequence will be an increased inclination among practitioners to use them. Survey methods, such as questionnaires and interviews/focus groups of users will be used to evaluate the quality of interaction and visualisation mechanisms. User studies and observational methods will be employed to understand users’ experiences of the systems to enable further refinement and implementation. Furthermore, crowd sourcing and citizen science can be used to ensure a large number of respondents, in order to make the assessment as scientifically robust as possible.

The role of the student will be to design and implement interaction mechanisms and visualisations of data arising from evolutionary computation. This will require generating experimental data that can be visualised from a range of optimisation problem types, using a variety of types of EA. Having developed these methods, they will be responsible for evaluating the use of their methods, and writing them up for presentation at leading international conferences and top journals in HCI and evolutionary algorithms.

For further information, please see: http://www.exeter.ac.uk/studying/funding/award/?id=2933

Best regards,
David

--
Dr. David Walker
Postdoctoral Research Fellow, Computer Science & Centre for Water Systems
College of Engineering, Mathematics & Physical Sciences
University of Exeter