Horizon researcher wins a prestigious machine learning contest

Dr Duncan Barrack, researcher in Horizon, has won a machine learning contest organised by a consortium of neural engineering institutions including the French INRIA research institute.

There were over 250 entrants from across the world and first prize for Duncan is $500, a slot to present the results
at the forthcoming IEEE Neural Engineering Conference, and an invitation to submit a paper to the IEEE Transactions on Biomedical Engineering.

The aim of the contest was to develop a machine learning model to predict when a brain computer interface ‘P-300 speller’ has made a mistake. A P-300 speller attempts to use brain waves collected via an Electroencephalography (EEG) cap placed on a user to spell words, and is designed to help enable people with locked-in syndrome to communicate.

The competition was hosted on the data science platform Kaggle, and more information on the challenge and the supporting institutions can be found at: https://www.kaggle.com/c/inria-bci-challenge