具有二维状态转移结构的随机逻辑及其在神经网络中的应用

季渊; 陈文栋; 冉峰; 张金艺; DavidLILJA

doi:10.11999/JEIT151233

具有二维状态转移结构的随机逻辑及其在神经网络中的应用

doi: 10.11999/JEIT151233

2.
(上海大学微电子研究与开发中心上海 200072) ②(上海大学机电工程与自动化学院上海 200072) ③(明尼苏达大学电子与计算机工程系明尼阿波利斯 55455)

基金项目:

国家自然科学基金(61376028)

计量
- 文章访问数: 1206
- HTML全文浏览量: 93
- PDF下载量: 544
- 被引次数: 0
出版历程
- 收稿日期: 2015-11-03
- 修回日期: 2016-04-08
- 刊出日期: 2016-08-19

Stochastic Logics with Two-dimensional State Transfer Structure and Its Application in the Artificial Neural Network

2.
(Microelectronic Research and Development Center, Shanghai University, Shanghai 200072, China)

Funds:

The National Natural Science Foundation of China (61376028)

摘要

摘要: 随机计算是一种特殊的基于概率数据码流的数学计算方法，其优点在于可以采用非常简单的数字逻辑完成复杂数学运算，从而大幅降低硬件实现成本。该文首先讨论了随机计算的基本原理和主要运算逻辑，论述了传统线性状态机的不足，并分析了一种2维状态转移拓扑结构，推导了通过2维有限状态机实现高斯函数的方法。在此基础上，提出一种随机径向基函数神经网络模型，其硬件实现成本非常低，而性能与传统神经网络相当。两类模式识别实验结果显示，所提出的随机径向基函数神经网络的输出值均方误差与相应结构传统神经网络的差别小于1.3%。FPGA实验结果显示，数据宽度为12位时，随机中间神经元的电路面积仅为传统插值查表结构的1.2%、坐标旋转数字计算方法(CORDIC)的2%。通过改变输入码流长度，该神经网络可以在处理速度、功耗和准确性之间作出平衡，具有应用灵活性，适用于对成本、功耗要求较高的应用如嵌入式、便携式、穿戴式设备。
- 随机计算 /
- 人工神经网络 /
- 径向基函数 /
- 模式识别
Abstract: Stochastic computing is a special algorithm that performs mathematical operations with probabilistic values of bit streams rather than traditional deterministic values. The main advantage of stochastic computing is its great simplicity of hardware arithmetic units for mathematical operations to reduce the circuit cost. This paper discusses the principle of the stochastic computing and its main arithmetic logic. It analyzes a two-dimension state transition topology structure, and discusses the Gaussian function implementation method based on the two-dimension Finite State Machin (FSM). Then, a low cost stochastic radial basis function neural network model is proposed. Results from two pattern recognition tests show that the difference of the mean squared error between the stochastic network output value and the corresponding deterministic network output value can be less than 1.3%. FPGA implementation results show that the hardware resource requirement of the proposed stochastic hidden neuron is only 1.2% of the corresponding deterministic hidden neuron with the interpolated look-up table, and is 2.0% of the CORDIC algorithm. The accuracy, speed and power of the stochastic network can be tradeoff dynamically. This network is suitable for the low cost and low power applications like embedded, portable and wearable devices.
- Stochastic computing /
- Artificial neural network /
- Radial Basis Function (RBF) /
- Pattern recognition

HTML全文

参考文献(20)

GAINES B R. Stochastic Computing Systems (Chapters) in Advances in Information Systems Science[M]. New York: Plenum, 1969: 37-172.

HAYES J P. Introduction to stochastic computing and its challenges[C]. 2015 52nd ACM/EDAC/IEEE Design Automation Conference (DAC), San Francisco, CA, USA, 2015: 1-3. doi: 10.1145/2744769.2747932.

ALAGHI A and HAYES J P. Survey of stochastic computing[J]. ACM Transactions on Embedded Computing Systems, 2013, 12(2s): 1-19. doi: 10.1145/2465787.2465794.

MOONS B and VERHELST M. Energy-efficiency and accuracy of stochastic computing circuits in emerging technologies[J]. IEEE Journal on Emerging and Selected Topics in Circuits and Systems, 2014, 4(4): 475-486. doi: 10.1109/JETCAS.2014.2361070.

BROWN B D and CARD H C. Stochastic neural computation. I. Computational elements[J]. IEEE Transactions on Computers, 2001, 50(9): 891-905. doi: 10.1109/12.954505.

QIAN Weikang, LI Xin, RIEDEL M D, et al. An architecture for fault-tolerant computation with stochastic logic[J]. IEEE Transactions on Computers, 2011, 60(1): 93-105. doi: 10.1109/TC.2010.202.

HAN Jie, CHEN Hao, LIANG Jinghang, et al. A stochastic computational approach for accurate and efficient reliability evaluation[J]. IEEE Transactions on Computers, 2014, 63(6): 1336-1350. doi: 10.1109/TC.2012.276.

ALAWAD M and LIN Mingjie. FIR filter based on stochastic computing with reconfigurable digital fabric[C]. 2015 IEEE 23rd Annual International Symposium on Field- Programmable Custom Computing Machines (FCCM), Vancouver, BC, Canada, 2015: 92-95. doi: 10.1109/FCCM. 2015.32.

TEHRANI S S, NADERI A, KAMENDJE G A, et al. Majority-based tracking forecast Memories for Stochastic LDPC Decoding[J]. IEEE Transactions on Signal Processing, 2010, 58(9): 4883-4896. doi: 10.1109/TSP.2010.2051434.

LI Peng, LILJA D J, QIAN Weikang, et al. Computation on stochastic bit streams digital image processing case studies[J]. IEEE Transactions on Very Large Scale Integration (VLSI) Systems, 2014, 22(3): 449-462. doi: 10.1109/TVLSI.2013. 2247429.

ZHANG Da and LI Hui. A stochastic-based FPGA controller for an induction motor drive with integrated neural network algorithms[J]. IEEE Transactions on Industrial Electronics, 2008, 55(2): 551-561. doi: 10.1109/TIE.2007.911946.

王守觉, 李兆洲, 陈向东, 等. 通用神经网络硬件中神经元基本数学模型的讨论[J]. 电子学报, 2001, 29(5): 576-580.

WANG Shoujue, LI Zhaozhou, CHEN Xiangdong, et al. Discussion on the basic mathematical models of neurons in general purpose neurocomputer[J]. Acta Electronica Sinica, 2001, 29(5): 576-580.

吴大鹏, 赵莹, 熊余, 等. 基于小波神经网络的告警信息相关性挖掘策略[J]. 电子与信息学报, 2014, 36(10): 2379-2384. doi: 10.3724/SP.J.1146. 2013.01701.

WU Dapeng, ZHAO Ying, XIONG Yu, et al. Alarm information relevance mining mechanism based on wavelet neural network[J]. Journal of Electronics Information Technology, 2014, 36(10): 2379-2384. doi: 10.3724/SP.J.1146. 2013.01701.

BROWN B D and CARD H C. Stochastic neural computation. II. Soft competitive learning[J]. IEEE Transactions on Computers, 2001, 50(9): 906-920. doi: 10.1109/12.954506.

LI Peng, LILJA D J, QIAN W K, et al. The synthesis of complex arithmetic computation on stochastic bit streams using sequential logic[C]. 2012 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), San Jose, CA, USA, 2012: 480-487. doi: 10.1145/2429384.2429483.

JI Yuan, RAN Feng, MA Cong, et al. A hardware implementation of a radial basis function neural network using stochastic logic[C]. 2015 Design, Automation Test in Europe Conference Exhibition (DATE), Grenoble, France, 2015: 880-883.

马承光, 仲顺安, LILJA D J, 等. 基于超几何分解的随机运算系统分析方法[J]. 电子与信息学报, 2013, 35(2): 355-360. doi: 10.3724/SP.J.1146.2012.00711.

MA Chengguang, ZHONG Shunan, LILJA D J, et al. Analysis method of stochastic computing system based on hypergeometric decomposition[J]. Journal of Electronics Information Technology, 2013, 35(2): 355-360. doi: 10.3724/ SP.J.1146.2012.00711.

施引文献

资源附件(0)

访问统计