The NCRG as part of the Mathematics group is looking for highly motivated individuals to fill PhD studentships in research areas covered by the group. The group has an internationally renowned research effort spanning theory and applications of pattern analysis, statistical physics, complex systems, stochastic nonlinear equations, and machine learning. Information on applying for postgraduate study at Aston can be found here.
We are currently inviting applications for the following projects:
NCRG graduate profiles
Listen to the following students discuss their experiences of studying in the NCRG. To view a video, click on the correponding image.
Current research projects Quality Assurance of Observational Data: A Bayesian Data Assimilation Approach
Supervisor: Dr. Dan Cornford
Observations of the current state of the atmosphere are critical to constructing accurate weather forecasts, and developing accurate climatologies. The Met Office currently makes use of several hundreds of thousands of observations each day to provide optimal initial conditions for its weather forecasts, in a process known as data assimilation. Collecting these measurements is expensive. One area of opportunity for improving the set of observations available at low cost is the use of user contributed and external sensor network data. The use of this data requires better quantification of the data quality, particularly biases. This project will develop statistical methods based on Bayesian inference to allow the assessment of data quality, using reference standards (data of known quality) and models for the spatial and temporal variation of the observed properties as well as the associated biases and uncertainties. In particular we will develop spatial models for surface air temperature, snow depth, wind speed and precipitation using and extending Gaussian process models. One of the more interesting challenges we will address during the project is how to best use the information available from the increasingly high resolution numerical weather prediction models being used within the Met Office. These models are key parts of the Met Office forecast system and this project will contribute to better understanding and characterising their uncertainties. This will enable their output to be used with greater confidence. The student would gain skills in probabilistic modelling, data assimilation and meteorology. For further information, see the project details here
. Topographic Information Visualisation
Supervisor: Prof. David Lowe
This project focuses on the mathematical problem of the relative dissimilarity representation of signals, driven by the real-world sonar array problem domain, and specifically aimed at massive data reduction whilst retaining informative content for the human operators. In response to the requirement for human decision support aids to reduce information overload and so to represent high dimensional structured data (data with interpoint relative similarity measures), this project aims to research prototype topographic visualisation models and compare and contrast with other state of the art models, and develop extended versions using models for uncertainty in data, structures and knowledge appropriate to the sonar sensor array domain. In addition, the project will explore the development of new types of data analysis and representation based on the dissimilarity representation. Nonlinearity and uncertainty in data and parameters and models will be included which will modify the visualisation space distributions. Dynamics and nonstationarity of the signal domain will be also be incorporated through feature extraction and adaptive models. For further information, see the project details here