Machine Learning for Neuroinformatics
By Leon French, UBC Bioinformatics
Abstract:
I will start with a quick introduction to neuroinformatics. A good
example is my work at the genome and connectome scales with large
datasets that require computational methods. I will present a few of
these interesting neuroinformatics datasets. With simple methods we
extracted several global relationships that provide new insight into
the rodent brain. We hope more complex methods from machine learning
and computational analysis can reveal stronger patterns.
Unfortunately, neuroinformatics research is limited by incomplete and
fragmented datasets. I describe my work that applies natural language
processing to formalize neuroscience literature. I present methods and
results describing the recognition and normalization of brain region
mentions. I have just recently completed those subtasks and will
present my early efforts to apply graph kernel based learning methods.