Publication

Direct Feedback Alignment Scales to Modern Deep Learning Tasks and Architectures