Variational Methods for Bayesian Independent Component Analysis
Data and information are, respectively, the resource and commodity of the modern world. Data is extracted from the world like crude oil from a reservoir - in its rawest form, data is of little use.
It is only when information is extracted via processing - like petroleum from oil - that data becomes `meaningful'. The discovery, analysis and recognition of patterns is of paramount importance if information is to be extracted from the raw observations on the world.
This is something that comes naturally to humans. The human brain is a pattern recogniser par excellance and, as such, mathematical pattern recognition forms the foundation of `intelligent' data processing by computers.
The mathematical study of patterns is rapidly becoming an area of huge significance, particularly as the world becomes increasingly data-driven and information-hungry. New novel and anthropogenic methods will be needed to interrogate, process, organise, store, access and understand this data quickly and intuitively.