Paper Info
Title
Softmax Regression Design for Stochastic Computing Based Deep Convolutional Neural Networks.
Abstract
Recently, Deep Convolutional Neural Networks (DCNNs) have made tremendous advances, achieving close to or even better accuracy than human-level perception in various tasks. Stochastic Computing (SC), as an alternate to the conventional binary computing paradigm, has the potential to enable massively parallel and highly scalable hardware implementations of DCNNs. In this paper, we design and optimize the SC based Softmax Regression function. Experiment results show that compared with a binary SR, the proposed SC-SR under longer bit stream can reach the same level of accuracy with the improvement of 295X, 62X, 2617X in terms of power, area and energy, respectively. Binary SR is suggested for future DCNNs with short bit stream length input whereas SC-SR is recommended for longer bit stream.
Year
DOI
Venue
2017
10.1145/3060403.3060467
ACM Great Lakes Symposium on VLSI
Field
DocType
Citations 
Softmax function,Convolutional neural network,Massively parallel,Computer science,Algorithm,Real-time computing,Artificial intelligence,Deep learning,Bitstream,Stochastic computing,Binary number,Scalability
Conference
5
PageRank 
References 
Authors
0.42
8
9
Name
Order
Citations
PageRank
Zihao Yuan190.85
Ji Li29710.87
Qinru Qiu31120102.58
Qinru Qiu41120102.58
Caiwen Ding514226.52
Ao Ren69611.53
Bo Yuan726228.64
Jeff Draper829826.31
Yanzhi Wang91082136.11