.

Research degrees

Inference

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.

Student Interview: Jack Raymond
Interview with Jack Raymond, PhD graduate.
 


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.

Using next generation computers and algorithms for modelling the dynamics of large biomolecular systems
Supervisor: Dr. Dmitry Nerukh

One of the main challenges in simulation of complex molecular processes such as protein folding or ligand-protein binding is the modelling of water: behaving as a structureless continuum in the bulk it needs to be represented at the atomistic level in a relatively small ‘core’ area of the system. The simulation of atomistic (explicit) water takes up to 90% of computing resources and makes the calculation prohibitively expensive. It is intuitively clear that the atomistic details are unnecessary in the areas distant from the biomolecule. However, in the vicinity of the biomolecule some water molecules are known to contribute to the biomolecular process in a very non-trivial way and their explicit modelling is decisively important.  The most natural representation of water in the bulk is provided by continuum hydrodynamics (CH).  Computer modelling in both representations, MD and CH, is well developed, but separated by a gap in the time and space scales accessible to simulations. Closing this gap is only possible if two directions are developed coherently: 1) new generation hardware, currently approaching the CH scales in MD simulations and 2) theory and software correctly joining the MD and CH representations.

The aims of this project are: to develop a new efficient theoretical and computational framework for hybrid MD-CH simulation of bio-chemically important processes at realistic time and space scales; to implement and test this framework in the world fastest supercomputing facility; and to conduct large scale simulations of trimethoprim (TMP) binding to dihydrofolate reductase (DHFR) and compare the predicted kinetic properties, the binding rate, with measured experimental values.

The project is a well balanced combination of state of the art computer hardware development, advanced numerical modelling (the triad of molecular dynamics, continuum fluid mechanics, and numerical methods) and cutting edge investigation of a biomedically important molecular system. For further information, see the project details here.

Employable Graduates; Exploitable Research