一种可应用于嵌入式系统的低复杂度超限学习机训练方法

doi:10.13190/j.jbupt.2017-110

北京邮电大学学报 ›› 2018, Vol. 41 ›› Issue (2): 9-14.doi: 10.13190/j.jbupt.2017-110

一种可应用于嵌入式系统的低复杂度超限学习机训练方法

张克终^1,2, 徐力³, 尉志青¹, 黄赛¹, 冯志勇¹

1. 北京邮电大学信息与通信工程学院, 北京 100876;
2. 北京工业大学北京未来网络科技高精尖创新中心, 北京 100124;
3. 中国科学院计算技术研究所, 北京 100190

收稿日期:2017-06-18 出版日期:2018-04-28 发布日期:2018-03-17
作者简介:张克终(1989-),男,博士生,E-mail:k.z.zhang_cn@ieee.org;冯志勇(1971-),女,教授,博士生导师.
基金资助:
国家自然科学基金项目（61631003，61525101）；深圳市海外高层次人才创新创业专项资金"孔雀团队"项目（KQTD2015033114415450）

Low Complexity Training Strategy for Extreme Learning Machine Used in Embedded System

ZHANG Ke-zhong^1,2, XU Li³, WEI Zhi-qing¹, HUANG Sai¹, FENG Zhi-yong¹

1. School of Information and Communication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China;
2. Beijing Advanced Innovation Center for Future Internet Technology, Beijing University of Technology, Beijing 100124, China;
3. Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China

Received:2017-06-18 Online:2018-04-28 Published:2018-03-17

摘要/Abstract

摘要： 为解决超限学习机复杂度较高的问题，提出了一种新型的超限学习机更新策略，称为序列超限学习机.避免了复杂度较高的逆矩阵运算，而且能够应用于嵌入式系统中.序列超限学习机比各种广泛应用的机器学习分类器具有更低的计算复杂度.基于实际数据集的仿真结果表明，序列超限学习机的分类精度比传统超限学习机和其他广泛应用的分类器更高，而且具有更短的训练时间.

关键词: 超限学习机, 神经网络, 分类算法, 低复杂度, 嵌入式系统

Abstract: Extreme learning machine (ELM) achieves faster training speed and higher classification accuracy, compared with other widely used classifiers, such as back propagation (BP), support vector machine (SVM), spectral clustering (SC), and so forth. However, ELM suffers from some drawbacks:1) ELM utilizes the calculation of inverse matrix for training, which cannot be adopted in the embedded system; 2) the training time of ELM increases dramatically for large-scale applications. To solve these drawbacks of ELM, a new training strategy called Sequential ELM (SELM) was proposed, which avoids the calculation of inverse matrix. Therefore, SELM can be adopted in the embedded system. It is proven that SELM achieves lower complexity than other widely used algorithms. Furthermore, simulations based on practical datasets indicate that the classification accuracy of SELM is higher than traditional ELM and other widely used classifiers with shorter training time.

Key words: extreme learning machine, neural network, classification, low complexity, embedded system

中图分类号:

TP183

张克终, 徐力, 尉志青, 黄赛, 冯志勇. 一种可应用于嵌入式系统的低复杂度超限学习机训练方法[J]. 北京邮电大学学报, 2018, 41(2): 9-14.

ZHANG Ke-zhong, XU Li, WEI Zhi-qing, HUANG Sai, FENG Zhi-yong. Low Complexity Training Strategy for Extreme Learning Machine Used in Embedded System[J]. JOURNAL OF BEIJING UNIVERSITY OF POSTS AND TELECOM, 2018, 41(2): 9-14.

参考文献

[1] Tang Jiexiong, Deng Chenwei, Huang Guangbin. Extreme learning machine for multilayer perceptron[J]. IEEE Transactions on Neural Networks and Learning Systems, 2016, 27(4):809-821.
[2] Chakrabarty A, Buzzard G T, Z·ak S H. Output-tracking quantized explicit nonlinear model predictive control using multiclass support vector machines[J]. IEEE Transactions on Industrial Electronics, 2017, 64(5):4130-4138.
[3] Yang Yang, Shen Fumin, Huang Zi, et al. Discrete nonnegative spectral clustering[J]. IEEE Transactions on Knowledge and Data Engineering, 2017, 29(9):1834-1845.
[4] Huang Gao, Huang Guangbin, Song Shiji, et al. Trends in extreme learning machines:a review[J]. Neural Networks, 2015, 61(Supplement C):32-48.
[5] Liu Tianchi, Yang Yan, Huang Guangbin, et al. Driver distraction detection using semi-supervised machine learning[J]. IEEE Transactions on Intelligent Transportation Systems, 2016, 17(4):1108-1120.
[6] Jin Bo, Jing Zhongliang, Zhao Haitao. Incremental and decremental extreme learning machine based on generalized inverse[J]. IEEE Access, 2017(5):20852-20865.
[7] Zhang Lei, Zhang David. Evolutionary cost-sensitive extreme learning machine[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017, 28(12):3045-3060.
[8] 刘渭滨, 邹智元, 邢薇薇. 模式分类中的特征融合方法[J]. 北京邮电大学学报, 2017, 40(4):1-8. Liu Weibin, Zou Zhiyuan, Xing Weiwei. Feature fusion methods in pattern classification[J]. Journal of Beijing University of Posts and Telecommunications, 2017, 40(4):1-8.
[9] Huang Kaizhu, Jiang Haochuan, Zhang Xuyao. Field support vector machines[J]. IEEE Transactions on Emerging Topics in Computational Intelligence, 2017, 1(6):454-463.
[10] Kuang Kelvin, Lee Chi, Chen Chiaoen. An eigen-based approach for enhancing matrix inversion approximation in massive MIMO systems[J]. IEEE Transactions on Vehicular Technology, 2017, 66(6):5480-5484.
[11] Zhang Xianchao, Zong Linlin, Liu Xinyue, et al. Constrained clustering with nonnegative matrix factorization[J]. IEEE Transactions on Neural Networks and Learning Systems, 2016, 27(7):1514-1526.

[1]	杨丽花, 聂倩, 呼博, 江婷. 联合BP神经网络与基扩展模型的信道预测算法 [J]. 北京邮电大学学报, 2023, 46(3): 13-18.
[2]	吴进, 赵博, 汶恒, 王宇. 一种异构架构的图神经网络加速器优化研究一种异构架构的图神经网络加速器优化研究[J]. 北京邮电大学学报, 2023, 46(3): 19-24.
[3]	王延凯龙华邵玉斌杜庆治王瑶. 基于融合特征MGCC的语种识别方法[J]. 北京邮电大学学报, 2023, 46(2): 116-121.
[4]	张昊阁邵玉斌龙华杜庆治周大春. 基于Gammatone尺度功率规整系数谱图的语种识别[J]. 北京邮电大学学报, 2023, 46(2): 122-128.
[5]	张昊阁邵玉斌龙华彭艺周大春. 基于对数Gammatone滤波器能量谱图的语种识别[J]. 北京邮电大学学报, 2023, 46(1): 38-43.
[6]	张天骐汪锐安泽亮方竹王雪怡. 基于多任务学习的MIMO-OFDM信号联合信噪比估计与调制识别[J]. 北京邮电大学学报, 2022, 45(6): 98-104.
[7]	余佳润王亚峰. 一种低信噪比下基于深度学习的DoA估计方法[J]. 北京邮电大学学报, 2022, 45(6): 119-125.
[8]	叶康张淑军郭淇李辉崔雪红. 基于CM-Transformer的连续手语识别[J]. 北京邮电大学学报, 2022, 45(5): 49-53,78.
[9]	武兴坤, 罗涛, 刘爱军, 杨明, 张文静. 面向主动脉缩窄辅助诊断的神经网络模型设计[J]. 北京邮电大学学报, 2022, 45(4): 1-6.
[10]	武兴坤罗涛刘爱军杨明张文静李剑峰李文秀. 面向主动脉缩窄辅助诊断的神经网络模型设计[J]. 北京邮电大学学报, 2022, 45(4): 1-6.
[11]	火元莲, 巩琪, 齐永锋, 安娅琦. 基于归一化最小均方算法的自适应核RBFNN[J]. 北京邮电大学学报, 2022, 45(2): 29-35.
[12]	许同乐, 孟良, 孔晓佳, 苏元浩, 孙砚飞. 基于EEMD的ICNN故障诊断方法[J]. 北京邮电大学学报, 2022, 45(2): 110-116.
[13]	李晓辉, 王先文, 樊韬, 刘佳文, 万宏杰. 基于卷积神经网络的高性能5G下行同步算法[J]. 北京邮电大学学报, 2022, 45(2): 117-123.
[14]	王靖, 姜竹青, 门爱东, 郭晓强, 王智康. 阶梯式图像去噪方法[J]. 北京邮电大学学报, 2022, 45(1): 52-57.
[15]	刘永利, 朱亚孟, 晁浩. 多策略MRFO算法的卷积神经网络超参数优化[J]. 北京邮电大学学报, 2021, 44(6): 83-88,95.

一种可应用于嵌入式系统的低复杂度超限学习机训练方法

Low Complexity Training Strategy for Extreme Learning Machine Used in Embedded System

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价