二进神经网络中的汉明球突及其线性可分性

杨娟; 陆阳; 黄镇谨; 王强

doi:10.3724/SP.J.1004.2011.00737

二进神经网络中的汉明球突及其线性可分性

doi: 10.3724/SP.J.1004.2011.00737

1.
合肥工业大学计算机与信息学院合肥 230009

计量
- 文章访问数: 2277
- HTML全文浏览量: 60
- PDF下载量: 796
- 被引次数: 0
出版历程
- 收稿日期: 2010-12-20
- 修回日期: 2011-02-18
- 刊出日期: 2011-06-20

Hamming Sphere Dimple in Binary Neural Networks and Its Linear Separability

1.
School of Computer and Infomation, Hefei University of Technology, Hefei 230009

摘要

摘要: 对于二进神经网络,剖析其神经元的逻辑意义对网络的规则提取是十分重要的, 而目前每个神经元所表达的线性结构的逻辑意义仍没有完全解决, 一部分线性函数的结构及其逻辑意义尚不明确. 本文在寻找线性可分结构的过程中,提出了汉明球突的概念, 给出其是否线性可分的判定方法,并得到二进神经元与线性可分的汉明球突等价的充要条件, 从而建立了判别线性可分的汉明球突的一般方法,并通过实例验证了该方法的有效性.
- 二进神经网络 /
- 线性可分函数 /
- 汉明球突 /
- 线性分类
Abstract: It is very important to analyze the logical meaning of neurons for extracting rules from binary neural networks (BNNs). However, the problem is that the logical meaning of linear structure expressed by neurons has not been completely solved and the structure of several linear functions is not clear. To solve this problem, we define a special structure called Hamming sphere dimple and provide the judgment method for linear separability. Furthermore, we obtain the necessary and sufficient condition for the equivalence between the linearly separable Hamming sphere dimple (LSHSD) and the binary neurons. Finally, we propose a general method for judging the LSHSD. This method is validated to be effective to judge the LSHSD through examples.
- Binary neural networks /
- linearly separable function /
- Hamming sphere dimple /
- linear classification

HTML全文

参考文献(1)

[1]

Lu Yang, Han Jiang-Hong, Gao Jun. Research on the minimal upper bound of the number of hidden nodes in binary neural networks. Pattern Recognition and Artificial Intelligence, 2000, 13(3): 254-257(陆阳, 韩江洪, 高隽. 二进神经网络隐元数目最小上界研究. 模式识别与人工智能, 2000, 13(3): 254-257)[2] Chen F Y, Chen G R, He G L, Xu X B, He Q B. Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations. IEEE Transactions on Neural Networks, 2009, 20(10): 1645-1658[3] Chen F Y, Chen G R, He Q B, He G L, Xu X B. Universal perceptron and DNA-like learning algorithm for binary neural networks: non-LSBF implementation. IEEE Transactions on Neural Networks, 2009, 20(8): 1293-1301[4] Lu Y, Yang J, Wang Q, Huang Z J. The upper bound of the minimal number of hidden neurons for the parity problem by binary neural networks. Science China Information Sciences, to be published[5] Chua L O. CNN: a paradigm for complexity. Visions of Nonlinear Science in the 21st Century. Singapore: World Scientific, 1999[6] Chen F Y, He G L, Chen G R. Realization of boolean functions via CNN: mathematical theory, LSBF and template design. IEEE Transactions on Circuits and Systems I: Regular Papers, 2006, 53(10): 2203-2213[7] Lu Yang, Han Jiang-Hong, Zhang Wei-Yong. Logical relation determination criteria and equivalence rule extraction on binary neural networks. Pattern Recognition and Artificial Intelligence, 2001, 14(2): 171-176(陆阳, 韩江洪, 张维勇. 二进神经网络逻辑关系判据及等价性规则提取. 模式识别与人工智能, 2001, 14(2): 171-176)[8] Lu Yang, Wei Zhen, Gao Jun, Han Jiang-Hong. Logical meaning of Hamming sphere and its general judgement method in binary neural networks. Journal of Computer Research and Development, 2002, 39(1): 79-86(陆阳, 魏臻, 高隽, 韩江洪. 二进神经网络中汉明球的逻辑意义及一般判别方法. 计算机研究与发展, 2002, 39(1): 79-86)[9] Lu Yang, Han Jiang-Hong, Wei Zhen. A general judging and constructing method of SP functions in binary neural networks. Acta Automatica Sinica, 2003, 29(2): 234-241[10] Ma Xiao-Min, Yang Yi-Xian, Zhang Zhao-Zhi. A new frame-work and some results for threshold function. Chinese Journal of Computers, 2000, 23(3): 225-230(马晓敏, 杨义先, 章照止. 一种新的阈值函数的分析框架及有关结论. 计算机学报, 2000, 23(3): 225-230)[11] Lu Yang, Han Jiang-Hong, Zhang Wei-Yong. Study of cartesian sphere in binary neural networks. Pattern Recognition and Artificial Intelligence, 2004, 17(3): 368-373(陆阳, 韩江洪, 张维勇. 二进神经网络中笛卡尔球的研究. 模式识别与人工智能, 2004, 17(3): 368-373)[12] Wegener I. The Complexity of Boolean Functions. New York: Wiley, 1987[13] Crounse K R, Fung E L, Chua L O. Efficient implementation of neighborhood logic for cellular automata via the cellular neural network universal machine. IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications, 1997, 44(4): 355-361[14] Nemes L, Chua L O, Roska A T. Implementation of arbitrary boolean functions on a CNN universal machine. International Journal of Circuit Theory and Applications, 1998, 26(6): 593-610[15] Haykin S. Neural Networks: A Comprehensive Foundation (Second Edition). New Jersey: Prentice Hall, 1998[16] Negnevitsky M. Artificial Intelligence: A Guide to Intelligent Systems (Second Edition). New Jersey: Addision-Wesley, 2004[17] Hassoun M H. Fundamentals of Artificial Neural Networks. Massachusetts: The MIT Press, 1995[18] Gray D L, Michel A N. A training algorithm for binary feedforward neural networks. IEEE Transactions on Neural Networks, 1992, 3(2): 176-194[19] Hua Qiang, Zheng Qi-Lun. The Hamming-graph learning algorithm of BNN. Chinese Journal of Computer, 2001, 24(11): 1250-1255(华强, 郑启伦. 二进神经网络的汉明图学习算法. 计算机学报, 2001, 24(11): 1250-1255)[20] Kim J H, Park S K. The geometrical learning of binary neural networks. IEEE Transactions on Neural Networks, 1995, 6(1): 237-247

施引文献

资源附件(0)

访问统计