The National Data Science Bowl, a data science competition where the goal was to classify images of plankton, ran from 15/12/2014 to 16/03/2015. Our team, consisting of Jeroen Burms, Pieter Buteneers, Jonas Degrave, Sander Dieleman, Ira Korshunova, Lionel Pigou and Aäron van den Oord, achieved the first place out of 1049 contestants, claiming the 100 000 dollar prize! Read all about our approach to the problem on Sander's blog post.
This BCI challenge, running from 19/11/2014 to 24/02/2015, was proposed as part of the IEEE Neural Engineering Conference (NER2015). The goal was to detect error potentials, i.e. brain responses occuring when unexpected errors are observed, from EEG data. Our lab member Thibault Verhoeven reached 12th place out of the 260 competing teams.
The goal of the challenge was to predict seizures from intracrenial EEG recordings. This competition ran from 25/08/2014 untill 17/11/2014. Our master student Ira Korshunova fisished 10th out of 504 teams.
The aim of this challenge, running from 23/01/2014 to 20/03/2014 and organised as part of the European Conference on Computer Vision (ECCV2014), was to recognise and identify human gestures from a continuous RGB-depth data stream. Lab member Lionel Pigou (at that time still a master's student) finished 5th out of the 35 competing groups.
The Galaxy Zoo challenge started on 20/12/13 and finished on 04/04/2014. The goal of this competition was to predict how people would classify images of galaxies. Sander Dieleman managed to finish in first place, and made a nice write-up about his winning solution over on his blog post.