Understanding how the brain gives rise to intelligence is one of the most important challenges facing scientists today. A starting point for intelligence is the brain’s transformation of sensory inputs (e.g. patterns of light which reach the eye) into a representation that is useful for reasoning (e.g. discrete objects with properties like a category). The purpose of this challenge is to explore different predictive models of visual representation that can account for the responses of individual neurons.
Specifically, this year’s challenge is to predict the responses of single neurons from area V4, an intermediate area in primate visual cortex, to natural image patches. The nature of the representation that V4 neurons support is still under vigorous debate, so a successful model will provide both a benchmark for future models and a functional model which can be probed in silico by neuroscientists seeking to develop their intuitions into these enigmatic neurons.
The top 3 performers will win respectively $300, $200, and $100 dollars. We actively welcome competitors from all fields, not just neuroscience.
The competition is open now and is currently set to close the evening of January 22nd. To compete, all you need to do is enter a submission at https://www.kaggle.com/c/uwndc19. Please see the linked website for further details.