Date | Thursday 14 November 2019
Time | 13:30 – Note unusual time
Venue| Senior Common Room, Level 2 (2D17), Priory Road Complex
Connectionist models are enjoying a renaissance as applications for processing big data, under the moniker “deep networks”. What can neuroscientists learn from this new momentum in machine learning and AI research? I will discuss how deep learning can help pose new questions for neuroscientists, reframe old debates, and provide a useful computational vocabulary for understanding natural intelligence. I will give some examples of how deep networks can help us understand the utility of various representational and computational schemes in biological brains, as measured with multielectrode recordings or macroscopic imaging methods such as fMRI. I will give examples of where human and machine information processing converge and diverge, including continual learning, curriculum learning, and structure learning.
All Welcome | Tea, coffee and cakes will be served after the seminar.