Random feedback weights support learning in deep neural networks

Comments



from Hacker News http://bit.ly/1rBbn2y