您现在的位置: 首页> 研究主题> Hopfield网络

Hopfield网络

Hopfield网络的相关文献在1993年到2022年内共计216篇,主要集中在自动化技术、计算机技术、无线电电子学、电信技术、数学 等领域,其中期刊论文190篇、会议论文21篇、专利文献325016篇;相关期刊145种,包括天津大学学报、科学技术与工程、西安电子科技大学学报(自然科学版)等; 相关会议21种,包括2012河南省计算机大会暨学术年会、第十三届全国容错计算学术会议、第二届全国信号处理与应用学术会议等;Hopfield网络的相关文献由440位作者贡献,包括郭鹏、叶世伟、张军英等。

Hopfield网络—发文量

期刊论文>

论文:190 占比:0.06%

会议论文>

论文:21 占比:0.01%

专利文献>

论文:325016 占比:99.94%

总计:325227篇

Hopfield网络—发文趋势图

Hopfield网络

-研究学者

  • 郭鹏
  • 叶世伟
  • 张军英
  • 许进
  • 史忠植
  • 王文杰
  • 郑宏伟
  • 韩璞
  • 马琳
  • 戎晓剑
  • 期刊论文
  • 会议论文
  • 专利文献

搜索

排序:

年份

    • 张波
    • 摘要: 近年来深度学习理论再度中兴,在机器学习视觉识别和听觉识别领域应用日益广泛.玻尔兹曼机是比较典型的深度学习神经网络,其网络权值的训练算法有多种,比较经典的如对比离差(CD)算法.目前的算法无法精确取得网络热平衡状态的期望值,只能计算近似梯度值,同时算法运算量大、运行时间长.对此,提出了一种RBM权值计算的方法.首先将RBM等效成Hopfield网络,然后利用DHNN权值设计方法设计权值矩阵,最后将RBM权值求解问题转化为求DHNN权值矩阵特征值和特征向量问题,通过实例说明计算过程并给予数据反向验证该算法的正确性.
    • 李易达
    • 摘要: 文章提出基于Hopfield网络的水利工程软土地基处理技术对比方法,并采用该方法比较了排水固结法、化学固结法和人工材料加筋加固法的处理效果.模型的有效性评价结果表明,Hopfield网络在水利工程软土地基沉降的模拟中具有较高的有效性;基于Hopfield网络的模拟试验结果可知,化学固结法的处理效果相对最佳,而排水固结法相对最差.研究成果可以为水利工程软土地基处理技术的优化选择提供合理的判断依据和有效的分析工具.
    • 吴新杰; 何在刚; 李惠强; 郑静娜; 陈玲; 许超; 陈跃宁; 颜华
    • 摘要: In view of low image reconstruction quality of electrical capacitance tomography ( ECT ) , an image reconstruction algorithm of ECT based on multi-criteria of Hopfield network is put forward.Firstly, the basic principle of ECT image reconstruction and Hopfield network was analyzed.Secondly, according to the characteristics of the ECT image reconstruction, four criterion functions were determined,including image entropy, the error sum of squares between measuring capacitances and estimating capacitances, lo-cal heterogeneity function of reconstruction image, and total variation of reconstruction image.These four criteria functions were introduced to the energy function of Hopfield network,and thus the dynamic equa-tions of Hopfield networks were deduced.On this basis, the iterative algorithm for ECT image reconstruc-tion was gained.Finally, the proposed method was verified through simulation experiment.Simulation re-sults show that the error of image reconstruction and correlation coefficient are better than corresponding indicators obtained by LBP and Landweber iterative algorithm.This ECT image reconstruction method is proven to be highly effective and accurate.It also provides an effective method and means for ECT image reconstruction algorithm.%针对电容层析成像技术中图像重建质量较差的问题,提出一种基于多准则Hopfield网络模型的电容层析成像的改进图像重建算法。首先分析了ECT图像重建和Hopfield网络的基本原理,然后根据ECT图像重建的特点确定了4种准则函数:图像熵、测量电容和估计电容的误差平方和、重建图像的局部非均匀性函数和总变差,并将这4种准则函数引入Hopfield网络的能量函数中,由此推导出Hopfield网络的动态方程,在此基础上得到ECT图像重建迭代算法,最后通过仿真实验对所提方法进行了验证。仿真实验结果表明利用此方法获得的重建图像误差和相关系数比LBP算法和Landweber迭代算法得到的相应指标要好。由此可见,该方法是一种有效的、精确度较高的ECT图像重建方法。
    • 王世强; 姜子炎; 邢建春; 代允闯; 杨启亮
    • 摘要: 本文针对现有空调系统传感器故障检测方法在实时性、适应性等方面的不足,基于建筑无中心智能控制系统,提出了一种诊断传感器慢漂移故障的无中心算法。依据基本物理原理建立约束方程,检验各个变量测量值的一致性,并通过各相邻节点的相互协作,借鉴Hopfield网络中能量函数等相关分析,实现传感器故障检测。该算法无需构建中心节点,不必增加硬件或时序冗余,使数据在系统底层就能自组织地完成诊断,具有并行性和简单性的特点,方便工程实施。以冷冻水二次泵系统为例,对算法进行了验证,结果表明:该方法能实现单个或多个传感器故障的检测,使空调系统传感器故障检测问题得到更合理的解决。%Due to the poor instantaneity and adaptability, the current methods for sensors fault detection are still less applied in practical project. Based on the decentralized intelligent control system of building, a decentralized algorithm for the faults of slow drift in sensors is proposed. The basic idea of the method is to detect the consistency of variables under the equality and inequality constraints established through physic principle. In the decentralized control system, each sensor is fitted with a decentralized controller such that it become a sensor-agent, which can communicate with other sensors and thus operate collaboratively to check out the faulty nodes assited by Hopfield network. Unlike traditional centralized and distributed algorithms, this algorithm does not need to build a central node, regard⁃less of increasing hardware or time redundancy, with good parallelism and simplicity. In the case of secondary chilled water pump system, this method is verified effective to detect the fault of sensors through simulation result.
    • 王晓娟; 杨永昕
    • 摘要: 计算机的功能非常强大,在处理图片方面也具有很好的性质.手写体图片的研究,在考古等方面有着重要的作用.本文的手写体图片是经多数人书写,保证了样本的差异性.在图片识别处理时,选用了识别性能较强的离散Hopfield网络,并针对Hopfield网络的特点,对手写体图片的进行中心归一化处理的改进,提高了识别效率.
    • 阮秀凯; 唐震洲; 张耀举; 陈孝敬; 陈慧灵
    • 摘要: To solve the special issue of electrical adaptive blind equalization in wireless spatial diversity optical coherent receivers, a new blind detection algorithm of multi- value QAM signals using output-feedback- bias ( OFB ) type complex discrete- time continuous state ( DTCS ) Hopfield neural network was presented. The OFB will not change the traditional Hopfield model. The proposed OFB- DTCS Hopfield neural network can meet special requirement of the multi- valued signal detection which need enlarger the search space. The blind detection problem of multi- valued QAM signals was transformed into solving a quadratic optimization problem. How to map the cost function of this optimization problem to the energy function of OFB- DTCS Hopfield neural network was also shown. The proof, analysis and its constraints of the energy function were shown, respectively. A complex activation function to fit this special problem was discussed. Then a special connective matrix was constructed to ensure algorithm detect signals correctly. Finally, detailed simulation results and performance comparison with other algorithm were shown to demonstrate farther the effectiveness, superiority and shortage of this new algorithm.%为解决无线分集相干光接收机的自适应盲检测问题,提出了一种新的离散时间连续状态的网络输出反馈偏置型的复Hopfield神经网络用以解决多值QAM信号的盲检测问题。反馈电压偏置的引入即不脱离传统Hopfield模型,又能有效满足多值信号检测时所需的搜索空间变大的特殊要求。全文完成多值信号盲检测的优化问题构造和能量函数的映射,给出能量函数的证明、分析和它的约束条件,给出适用该问题的激活函数的基本特征,正确盲检测信号的权矩阵的配置方法。最后,通过详细的仿真结果展示和与其他算法性能对比进一步验证算法的有效性和优越性并指出算法所存在的问题和下一步的研究方向。
    • 王静; 唐向宏; 林新建
    • 摘要: Aiming at problems of tamper detection and recovery, a recoverable watermarking algorithm for text authentication and synonym replacement based on synonym replacement technology and associative memory function of Hopfield neural network is proposed. The text is divided into replaceable synonyms and non-replaceable words. The feature information of replaceable synonyms is extracted according to the position in their thesaurus and the feature information of non-replaceable words is extracted according to the structure and stroke of Chinese characters in the text. Then, the watermark is embedded by synonym replacement. The information of watermark and the feature information of non-replaceable words are input into the Hopfield neural network and they are trained to realize tamper detection and recovery function of replaceable synonyms. The simulation results show that the proposed algorithm has good robustness, tamper detection performance and recoverability, and by this algorithm, tamper detection and the location of replaceable synonyms and non-replaceable words are implemented to realize text authentication, recover the original replaceable synonyms and achieve recovery.%针对水印文本的篡改检测和恢复问题,利用Hopfield网络的联想记忆功能,提出一种Hopfield神经网络与同义词替换技术相结合的文本认证与同义词替换的可恢复水印算法。算法首先将文本分为可替换同义词和非替换词语,利用可替换同义词在其同义词库中的位置及汉字笔画和结构特征,分别提取文本可替换同义词的特征信息和非替换词语的特征信息。然后通过同义词替换实现水印嵌入,并将水印信息和非替换词语的特征信息输入Hopfield神经网络进行训练,实现篡改检测与可替换同义词的恢复功能。实验仿真表明,该算法具有较好的鲁棒性、篡改检测性和恢复能力,能篡改检测和定位可替换同义词、非替换词语,实现认证功能,且能恢复被替换的同义词,实现恢复功能。
    • 隗兵; 戴文战
    • 摘要: Hopfield网络容量大小对网络模式识别正确率有重要影响.为进一步提升Hopfield的网络容量,提出了一种基于克隆选择算法优化Hopfield网络容量的方法.首先将克隆选择算法引入到Hopfield网络中,以Hopfield网络的初始输入作为克隆选择算法中的抗原;然后随机产生权值矩阵作为克隆选择算法的初始抗体;最后依据克隆选择算法对初始抗体进行克隆、交叉、变异,根据亲和力的大小选择出网络的优化权值,以提升Hopfield网络容量.将上述方法应用于含噪声的样本识别,实验结果表明:与传统的Hopfield网络相比,所提出的方法能有效地提升Hopfield网络的容量.为提高Hopfield神经网络的记忆容量提供了一种新的思路.
    • 张俊亮; 何晓雪; 宋云霞
    • 摘要: Hopfield神经网络常被用于非线性约束优化的求解,属于动力学方法,可用来求解乘子法的子问题,且只需要进行一阶导数的计算.以逐渐衰减的高斯噪声信号为基础,对随机神经网络进行构建.初始温度能从很大程度上影响随机神经网络,随机神经网络也因此十分难以跳出局部极小值的限制,为有效解决该问题,神经网络在运行的过程中,对结合模拟退火的欧拉法进行了应用.运用此方法对喷气教练机进行总体优化设计,可知,算法具有数值稳定性高、求解精度高的特点,同时借助了由拉氏乘子提供的约束敏度信息,进行了相应的设计要求权衡,并对某型干线旅客机的机翼气动P结构综合设计问题进行了分析和研究.
    • 涂岩恺; 鄢煜尘
    • 摘要: 为了减小笔迹图像中的书写波动与噪声,研究了基于蚁群算法优化Hopfield神经网络的图像规范化预处理方法,使网络调整后的图像更接近标准样本,并对其进行特征提取和分类鉴别.同时引入动态组网的系统结构和网络联想自我评价方法,在大样本笔迹数据库上进行实验表明,该方法能够对笔迹图像中的复杂波动与噪声进行有效的规范化处理,以提高计算机笔迹鉴别的准确性,10候选鉴别正确率可提高到95.65%.
  • 查看更多

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号