I am a PhD student in machine learning at the Institute of Adaptive and Neural Computation, School of Informatics, University of Edinburgh, supervised by Dr. Amos Storkey and Dr. David McGonigle.
I am interested in learning within dynamical systems, and in particular continuous time systems. Much of my work addresses Bayesian filtering, smoothing and parameter estimation, most significantly using Monte Carlo techniques such as particle filtering. A side theme of the work is tackling complex, potentially large-scale problems with methods amicable to distributed computing.
At present this work is applied to revealing the latent neural activity underlying Functional Magnetic Resonance Imaging (fMRI), although broader applications potentially include financial, environmental and physical modelling.
Publications
- Murray, L.M. and Storkey, A.J. (2008) Continuous Time Particle Filtering for fMRI. To appear in Advances in Neural Information Processing Systems, 20. [full text] [poster]
- Storkey, A.J., Simonotto, E., Whalley, H., Lawrie, S., Murray, L.M. and McGonigle, D. (2007) Learning structural equation models for fMRI. Advances in Neural Information Processing Systems, 19, 1329-1336.
Software
The dysii Dynamic Systems Library and fmrii fMRI Modelling Library are both direct products of my research work to date. I try to keep most of the code I write general enough to be included in these two libraries.
Tint is a fun little side project spun off from a brainstorm session. It uses K-means clustering to detect the colour groups in an image, and provides for some artistic effects based on these.
