Network Architecture of Cortex and Cerebellum for Supporting Super-learning
Serhat Çağdaş, Yalova University, Turkey; Ismail Akturk, Ozyegin University, Turkey; N. Serap Şengör, İstanbul Technical University, Turkey
Posters 3 Poster
Pacific Ballroom H-O
Sat, 27 Aug, 19:30 - 21:30 Pacific Time (UTC -7)
Although there are plenty of evidence about the interplay between supervised, unsupervised and reinforcement learning mechanisms in the brain, generally these mechanisms are modeled in isolation. In this work, a bio-plausible network architecture is proposed to train a neo-cortical network in interaction with cerebellum. We aim to train the recurrent cortical network with a temporal pattern. Cerebellar network is given a role in control of the population activity of the cortex by a lateral inhibition mechanism. Therefore, it is expected that cerebellum with supervised learning will prevent the activity of cortex to explode or diminish.