Open Peer Review. Open Publishing. Open Access. Open Discussion. Open Directory. Open Recommendations. Open API. Open Source.
ResBinNet: Residual Binary Neural Network
Mohammad Ghasemzadeh, Mohammad Samragh, Farinaz Koushanfar
Feb 15, 2018 (modified: Feb 15, 2018)ICLR 2018 Conference Blind Submissionreaders: everyoneShow Bibtex
Abstract:Recent efforts on training light-weight binary neural networks offer promising execution/memory efficiency. This paper introduces ResBinNet, which is a composition of two interlinked methodologies aiming to address the slow convergence speed and limited accuracy of binary convolutional neural networks. The first method, called residual binarization, learns a multi-level binary representation for the features within a certain neural network layer. The second method, called temperature adjustment, gradually binarizes the weights of a particular layer. The two methods jointly learn a set of soft-binarized parameters that improve the convergence rate and accuracy of binary neural networks. We corroborate the applicability and scalability of ResBinNet by implementing a prototype hardware accelerator. The accelerator is reconfigurable in terms of the numerical precision of the binarized features, offering a trade-off between runtime and inference accuracy.
TL;DR:Residual Binary Neural Networks significantly improve the convergence rate and inference accuracy of the binary neural networks.
Keywords:Binary Neural Networks, Residual Binarization, Deep Learning
Enter your feedback below and we'll get back to you as soon as possible.