基于深度自动编码器的脑网络状态观测矩阵降维方法

针对人脑网络状态观测矩阵高维无特征的特点,给出了一种基于深度自动编码器(DAE)的降维算法.利用深度学习网络,将高维的人脑网络空间表达映射到低维的本质特征空间中,为进一步提炼脑网络的动态性能提供了基础.实验结果证明:应用该方法可以达到有效的降维效果,且降维后脑网络状态通过自组织特征映射聚类具有一定的规律性,从而为脑网络的动态特性研究提供了基础.     

基于动态膜计算系统的自组织网络广播算法

应用仿生学原理设计了一种新型的广播算法——基于动态膜计算系统的广播算法,提出了动态膜计算系统,并给出了解决自组织网络中广播问题的规则集。系统中节点转播信息的优先权通过节点间距离及邻居个数确定,适合不同密度的网络;并根据信息数与门限值的比较,取消部分节点的转播权,从而提高了可达率和转播节省率。仿真测试验证了此系统用于广播是可行的、高效的,为设计无线自组织网络中的广播算法提供了新思路。     

基于自结构动态递归模糊神经网络的无人机姿态控制*

针对无人机非线性、强耦合等特点,提出了基于该自结构动态递归模糊神经网络的姿态控制系统,给出了基于Lyapunov函数的系统稳定性证明。对四层模糊神经网络进行了优化和改进,设计了自结构动态递归模糊神经网络,该网络可以根据系统状态在线更新权值、创建/删除节点、优化网络结构。仿真表明：该控制方法的突出优点是,在兼顾考虑了系统中的不确定性因素、非线性因素及外部干扰并存的情况下,保证系统的稳定性和跟踪性能;同时此网络结构比固定结构的模糊神经网络响应速度快,因此更具优越性。     

促分裂原活化蛋白激酶激活的蛋白激酶-2全新抑制剂的3D-QSAR研究

促分裂原活化蛋白激酶激活的蛋白激酶-2(MK-2)可以作为治疗类风湿性关节炎(RA)的靶标,并显示出巨大的潜力,此类酶的抑制剂用于治疗RA具有相当好的疗效,一系列β咔啉衍生物对于MK-2具有高的活性和选择性。通过自组织分子场(SoMFA)分析方法建立了MK-2抑制剂———β咔啉衍生物的三维定量构效关系(3D-QSAR)模型。SoMFA模型的相关系数r2=0.731,交叉验证相关系数q2=0.686,标准偏差s=0.388,Fisher检验值F=63.538。该模型具有很好的稳定性和较为满意的预测能力。     

自组织与自组织城市

自组织理论和方法在当代城市地理学和规划理论的前沿领域有着非常重要而深刻的应用 ,这方面的研究工作已经有30多年的历史乃至更早。文章评介了自组织城市理论、模型和自组织城市规划思想与方法 ,解释了自组织概念的基本含义 ,分别介绍了自组织城市研究的 7个重要领域 :耗散城市、协同城市、混沌城市、分形城市、细胞城市、沙堆城市以及FACS和IRN城市 ,讨论了基于自组织研究的空间复杂性问题     

A Hierarchical Self-Organizing Map Model for Sequence Recognition

A novel neural model made up of two self-organizing maps nets – one on top of the other – is introduced and analysed experimentally. The model makes effective use of context information, and that enables it to perform sequence classification and discrimination efficiently. It was successfully applied to real sequences, taken from the third voice of the sixteenth four-part fugue in G minor of the Well-Tempered Clavier (vol. I) of J.S. Bach. The model has an application in domains which require pattern recognition, or more specifically, which demand the recognition of either a set of sequences of vectors in time or sub-sequences into a unique and large sequence of vectors in time.     

Energy-efficient cluster-based artificial intelligence routing for wireless sensor networks

Artificial intelligence (AI)-based wireless sensor network technology is the future of advancement for real-time applications. With AI wireless sensor network technology, it is possible to collect data from any environment, analyze in real time, and use it to optimize processes and operations. AI wireless sensor network technology provides an unprecedented level of accuracy as well as the ability to detect even the slightest changes in a given environment. The AI-based approach uses clustering-based techniques with self-organizing map (SOM) for energy conservation in resource-constrained networks. By clustering the network, it becomes more energy efficient, as data can be shared among members of a cluster without needing to be transmitted across multiple nodes. The proposed AI cluster-based routing approach outperforms in terms of energy consumption and computational challenges of the network. The results obtained demonstrate the proposed approach to achieve lower energy consumption than the existing algorithms while providing the same level of performance in terms of throughput and latency, as well as a comparison with traditional justification techniques.     

一种基于自组织神经网络的彩色图像分割方法

提出了1种基于自组织神经网络的彩色图像分割方法.它首先在空间-颜色联合域内利用自组织神经网络对图像进行聚类,再结合区域生长的方法和区域邻接列表,对图像中的相似颜色区域进行合并,并去除像素数量较少的小区域,从而得到最后的分割结果.在聚类后的图像上进行分割,可以较好的把1个物体对象只表示为1个或少数几个分割区域,明显减少对象物数目,较好地克服了传统区域生长方法的过分割问题.     

Fuzzy Clustering in Software Reusability

This paper is concerned with the organization and retrieval of reusable software components with the aid of unsupervised learning. The methods considered of unsupervised learning include FUZZY ISODATA and Kohonen self-organizing maps. The key issues addressed in the study include information retrieval in the presence of incomplete information, and domain specific enhancements of unsupervised learning, including those of partial supervision. The primary intention is to reveal how the learning mechanism can accommodate individual preferences (profile) of the users viewed as a significant component of organization and retrieval algorithms. Numerical examples use a set of MS-DOS system commands and a collection of reusable C++ classes. © 1997 by John Wiley & Sons, Ltd.     

基于P2P计算模式的自组织网络路由模型

通过使用peer-to-peer(P2P)计算模式在Internet物理拓扑基础上建立一个称为P2P覆盖网络(P2Poverlay network)的虚拟拓扑结构,有效地建立起一个基于Internet的完全分布式自组织网络路由模型--分级集中式自组织网络路由模型(hierarchical aggregation self-organizing network,简称HASN).分别描述了HASN路由模型的构建目标和体系结构,并详细分析了HASN采用的基于P2P计算模式的分布式命名、路由发现和更新算法HASN Scale,并在仿真实验的基础上,对HASN路由模型的性能进行了验证.