PhD Research Studentship in Robustness of Complex Systems, Brunel University
How do complex networks learn how to develop their architecture and also cope with disturbance or damage?
Complexity and robustness have become major challenges as systems become networked and hierarchical. This collaborative project between disciplines of engineering and psychology, aims to assess the extent to which the means (‘mechanisms') of learning in human and artificial intelligence, and also in reconfiguring wireless networks, are robust and universal.
Learning in humans and machines occurs amidst noise, and wireless networks have to maintain function for roving users amidst noisy environments. Therefore robustness has become a major challenge as complex systems become networked and hierarchical.
For example, wireless communications and networks have to maintain function for roving users amidst noisy environments - the future challenge will be to better accommodate changing dynamics of users and their mobility requiring reconfigurable and scalable networks using robust techniques.
Complexity characterizes learning in at least two different ways. First, there is the complexity of the material to learn, which is the external environment; second, there is the complexity of the superlative learning mechanisms and the internal representations used to code the information to learn. In the latter case, the question of robustness is of paramount importance. Given that the input is often noisy and learning systems are not totally reliable, the question arises as to whether, and if so how, the learning mechanisms are sufficiently robust.
The project supervisors are:
Dr Mark Atherton (Mechanical Engineering) for Mechanisms of robustness; Professor Fernand Gobet (Psychology) for Acquisition of syntax and vocabulary; Professor Hamed Al-Raweshidy (Electronic & Computer Engineering) for Reconfigurable mobile networks.
The project objectives are to deliver:
- A better understanding of the extent to which models using discrimination networks are robust to diverse types of damage. Beyond language acquisition, these models are used for simulating numerous phenomena, such as the acquisition of expertise and problem solving in artificial intelligence, so the research will have an important impact in cognitive science as well as mobile networks.
- Understanding of the robustness parameters of a scalable, reconfigurable network.
- Shed light on whether some disorders in language acquisition can be seen as dysfunctions in mechanisms normally ensuring robustness of learning.
- A taxonomy of universal robustness for learning in complexity science.
We are seeking candidates with at least a 2:1 honours degree in engineering, computer science or equivalent. A demonstrable ability in computing programming is essential and good mathematical and modelling skills are desirable. The student must be Home/EU and not have been previously registered for a research degree due to funding restrictions
Closing Date for applications received - Wednesday 18 June
Funding is available for a three year Home/EU PhD studentship based on £14600 stipend per annum and £3235 fees per annum. In addition, £1000 (total over the three year studentship) is available for the PhD student's project and travel costs.
For more information contact Dr Mark Atherton on 01895 266690 or mark.atherton@brunel.ac.uk before 18 June 2008.
Click here for Employer Profile
