一级黄色片免费播放|中国黄色视频播放片|日本三级a|可以直接考播黄片影视免费一级毛片

高級搜索

留言板

尊敬的讀者、作者、審稿人, 關(guān)于本刊的投稿、審稿、編輯和出版的任何問題, 您可以本頁添加留言。我們將盡快給您答復(fù)。謝謝您的支持!

姓名
郵箱
手機(jī)號碼
標(biāo)題
留言內(nèi)容
驗(yàn)證碼

高階神經(jīng)網(wǎng)絡(luò)連接權(quán)的稀疏化及其刪減算法

李守麗,  李金艷,  李望超

李守麗, 李金艷, 李望超. 高階神經(jīng)網(wǎng)絡(luò)連接權(quán)的稀疏化及其刪減算法[J]. 電子與信息學(xué)報, 1999, 21(2): 182-185.
引用本文: 李守麗, 李金艷, 李望超. 高階神經(jīng)網(wǎng)絡(luò)連接權(quán)的稀疏化及其刪減算法[J]. 電子與信息學(xué)報, 1999, 21(2): 182-185.
Li Shouli, Li Jinyan, Li Wangchao. SPARSED CONNECTION WEIGHTS OF HIGHER-ORDER NEURAL NETWORK AND ITS PRUNING ALGORITHM[J]. Journal of Electronics & Information Technology, 1999, 21(2): 182-185.
Citation: Li Shouli, Li Jinyan, Li Wangchao. SPARSED CONNECTION WEIGHTS OF HIGHER-ORDER NEURAL NETWORK AND ITS PRUNING ALGORITHM[J]. Journal of Electronics & Information Technology, 1999, 21(2): 182-185.

高階神經(jīng)網(wǎng)絡(luò)連接權(quán)的稀疏化及其刪減算法

SPARSED CONNECTION WEIGHTS OF HIGHER-ORDER NEURAL NETWORK AND ITS PRUNING ALGORITHM

  • 摘要: 本文首先研究完全連接型高階神經(jīng)網(wǎng)絡(luò)的逼近能力,并證明了定義在{0,1}N上的任意布爾函數(shù)都可以由完全連接的高階神經(jīng)網(wǎng)絡(luò)來實(shí)現(xiàn)。接著提出了旨在簡化網(wǎng)絡(luò)結(jié)構(gòu)的去除冗余連接權(quán)刪減算法,并用于高階神經(jīng)分類器的稀疏化實(shí)現(xiàn)。模擬實(shí)驗(yàn)結(jié)果證明了這種算法的有效性。
    Abstract: In this paper, the fully-connected higher-order neuron and sparsed higher-order neuron are introduced, the mapping capabilities of the fully-connected higher-order neural networks are investigated, and that arbitrary Boolean function defined from {0,1}N can be realized by fully-connected higher-order neural networks is proved. Based on this, in order to simplify the networks architecture, a pruning algorithm for eliminating the redundant connection weights is also proposed, which can be applied to the implementation of sparsed higher-order neural classifier. The simulated results show the effectiveness of the algorithm.
    • HTML全文
  • Paretto P, Niez J J. Long term memory storage capacity of multiconnected neural networks, Biol[J].Cybern.1986, 54(3):53-63[2]Baldi P. Neural networks, orientations of the supercube and algebraic threshold functions, IEEE Trans. on Inform. Theory, 1988, IT-34 (3): 523-530.[3]Giles C L, Maxwell T. Learning, invariance, and generalization in high-order neural networks, Applied Optics, 1987, 26(23): 4972-4978.[4]Giles C L, Chen D, Miller C B, et al. Second-order recurrent neural networks for grammatical[5]inference, Proc. Int. Joint Conf. Neural Networks, IJCNN91, Seattle: vo1.2, 1991, 273-281.Lion R, Azimi-Sadjadi M R, Dent R. Detection of dim targets in high cluttered background using high order correlation neural network, Proc. Int. Conf. Neural Networks, IJCNN91, Seattle: vol.1, 1991, 701-706.[6]Shun-ichi Amari. Dualistic geometry of the manifold of higher-order neurons, Neural Networks, 1991, 4(5): 443-451.[7]李金艷.稀疏高階神經(jīng)網(wǎng)絡(luò)的通近能力及其算法的研究:[博士論文].華南理工大學(xué),1996.[8]李金艷,余英林.二層高階神經(jīng)網(wǎng)絡(luò)對任意布爾函數(shù)的實(shí)現(xiàn),華南理工大學(xué)學(xué)報,1995, 23(10): 111-116.[9]Fahner G, Eckmiller R. Structural adaption of parsimonious higher-order neural classifiers, Neural Networks, 1994, 7(2): 279-289.
  • 加載中
計量
  • 文章訪問數(shù):  617
  • HTML全文瀏覽量:  117
  • PDF下載量:  858
  • 被引次數(shù): 0
出版歷程
  • 收稿日期:  1997-08-25
  • 修回日期:  1998-07-26
  • 刊出日期:  1999-03-19

目錄

    /

    返回文章
    返回