Incorporating context in text analysis by interactive activation with competition artificial neural networks

Many segments of modern society, including marketing, politics, government, activism and public safety, desire the ability to find relationships, thus meaning, in public discourse. This can be accomplished by analyzing communication documents according to their content. The increasing use of the Internet for public dialog has made Internet communication a potentially rich source of data in this regard. This study explores the use of an interactive activation with competition (IAC) artificial neural network (ANN) to find relationships in email texts. A modified fully recurrent IAC network was used to process 69 email messages from two threads in the Open Library/Information Science Education Forum using two variations of the self-organizing phase of network formation. These variations were: (1) with and (2) without a linear decay function applied between sentences to the external activation of nodes. The use of the linear decay function, which could be considered a method for including context, produced three positive effects: the entire network was more differentiated from keywords; the keywords were more highly associated with each other, and; roughly half the number of noise strings were highly associated with keywords.     

基于模体的科学家合作网络基元特征分析

基于复杂网络构建科学家合作网络并分析其全局结构特征,是理解科学家间合作机制的重要手段和研究热点。实际上,辨识其局部结构可以进一步增进对合作机制的理解。基于复杂网络局部结构研究的前沿理念和方法——模体,构建并辨析若干大型科学家合作网络的基元特征:模体特性、子图浓度以及子图自下而上的组合机制。研究表明不同领域科学家合作网络的模体与反模体形式具有共性特征,网络的子图相对浓度则因不同领域科学家之间合作强度及网络发展程度差异而具有不同的分布特征。此外,科学家合作网络具有相似的自下而上构建机制。研究提供了基于模体的科学家合作网络结构辨识方法。     

京津冀协同发展背景下区域高等教育均衡发展研究

京津冀协同发展是国家重大战略,而高等教育协同是其重要内容。京津冀高等教育协同发展在顶层设计、政府协同、高校合作等方面还缺乏系统性、整体性和互动性。推动京津冀高等教育协同发展,需要进一步加强顶层设计、资源疏解、合作共建、资源共享,推动区域高等教育均衡发展。     

Stability of memristor neural networks with delays operating in the flux-charge domain

The paper considers a class of neural networks where flux-controlled dynamic memristors are used in the neurons and finite concentrated delays are accounted for in the interconnections. Goal of the paper is to thoroughly analyze the nonlinear dynamics both in the flux-charge domain and in the current-voltage domain. In particular, a condition that is expressed in the form of a linear matrix inequality, and involves the interconnection matrix, the delayed interconnection matrix, and the memristor nonlinearity, is given ensuring that in the flux-charge domain the networks possess a unique globally exponentially stable equilibrium point. The same condition is shown to ensure exponential convergence of each trajectory toward an equilibrium point in the voltage-current domain. Moreover, when a steady state is reached, all voltages, currents and power in the networks vanish, while the memristors act as nonvolatile memories keeping the result of computation, i.e., the asymptotic values of fluxes. Differences with existing results on stability of other classes of delayed memristor neural networks, and potential advantages over traditional neural networks operating in the typical voltage-current domain, are discussed.     

基于RBF神经网络的物流业能源需求预测

随着中国经济的快速发展,物流业需求快速增长,规模不断扩大,也带来了能源消耗的增长.研究中国物流业能源消费水平以及能源需求,有利于物流业节能工作的开展,缓解能源压力.本文选取了影响物流业能源需求的11个主要因素,基于径向基神经网络对2001-2012年间中国物流业能源需求相关数据进行模拟与仿真,在此基础上对2016年和2020年物流业能源需求量进行了预测,并分析了11个影响因素的重要性和测算了物流业的能源效率.研究结果表明:①2001-2012年间中国物流业能源消耗总量在不断增加,随着物流业的进一步发展,到2020年物流业能源消费总量将达到51261.92万t标准煤;②在解决物流业能源需求预测问题时,RBF神经网络比GM(1,1)预测模型,BP神经网络方法有更高的预测精度;③通过RBF神经网络变量重要性分析发现固定资产投资对物流业能源消费量的影响程度最大;④目前物流业能源效率明显低于全国能源效率,为节约能源,提高能源利用效率,物流业需要转变能源利用方式和发展模式.     

Global h-synchronization of stochastic delayed high-order inertial neural networks subject to Markovian jump parameters

As a first exploration, this paper proposed the second-order response system (SORS) method to study the global h-synchronization (SGhS) for high-order stochastic delayed inertial neural networks subject to Markovian jumping parameters. Different from previous studies, this paper avoids reducing the order of the original drive system via substitution of variables, and directly gives the corresponding SORS. In the following, under the framework of the SORS method, a regulation function dependent Lyapunov–Krasovskii functional (RFD–LKF) is designed to realize that the considered dynamics are globally mean square h-synchronous. Particularly worth mentioning is that the method proposed in this paper can greatly reduce the amount of calculation and control cost. Ultimately, via a typical example, the superiority and validity of the derived h-synchronous criterion can be well demonstrated.     

Towards the minimum-cost control of target nodes in directed networks with linear dynamics

Determining an input matrix, i.e., locating predefined number of nodes (named “key nodes”) connected to external control sources that provide control signals, so as to minimize the cost of controlling a preselected subset of nodes (named “target nodes”) in directed networks is an outstanding issue. This problem arises especially in large natural and technological networks. To address this issue, we focus on directed networks with linear dynamics and propose an iterative method, termed as “L₀-norm constraint based projected gradient method” (LPGM) in which the input matrix B is involved as a matrix variable. By introducing a chain rule for matrix differentiation, the gradient of the cost function with respect to B can be derived. This allows us to search B by applying probabilistic projection operator between two spaces, i.e., a real valued matrix space R^N?×?M and a L₀ norm matrix space $R_{L_0}^{N\times M}$ by restricting the L₀ norm of B as a fixed value of M. Then, the nodes that correspond to the M nonzero elements of the obtained input matrix (denoted as $B^{L_0}$) are selected as M key nodes, and each external control source is connected to a single key node. Simulation examples in real-life networks are presented to verify the potential of the proposed method. An interesting phenomenon we uncovered is that generally the control cost of scale free (SF) networks is higher than Erdos-Renyi (ER) networks using the same number of external control sources to control the same size of target nodes of networks with the same network size and mean degree. This work will deepen the understanding of optimal target control problems and provide new insights to locate key nodes for achieving minimum-cost control of target nodes in directed networks.     

粒子群神经网络算法在水电工程投资估算中的应用

文章基于粒子群BP神经网络,同时结合水利水电工程等级划分标准建立大型水电工程投资估算模型。并利用MATLAB软件实现了模型的训练与测试。最后对已建水电工程进行了预测。结果表明模型改进了现有水电工程投资估算方法,在水电工程投资估算中具有较高的精度与泛化能力。     

Divide to connect: Reorganization through R&D unit spinout as linking context of intra-corporate networks

The paper examines the formation of clique-spanning ties in intra-corporate co-invention networks, by focusing on the impact of R&D unit spinouts. Using data on thirty-one spinouts in eight large U.S. information and communication technology corporations, we show that the reorganization of R&D units through corporate spinouts is associated with an increase in the extent to which inventors employed in the unit collaborate with inventors located in other cliques within the corporate co-invention network. Interestingly, the spinout effect spills over to all members of the clique of spun-out inventors, also including those who remain with the parent firm. The interpretation of these empirical findings, grounded on existing theories and on the views of inventors and executives involved in the reorganization events, suggests that corporate spinouts generate a shock in intra-corporate research collaboration dynamics, which loosens clique lock-in effects and contributes to reset cliques’ boundaries in the intra-corporate research network.     

A modified Lyapunov functional with application to stability of neutral-type neural networks with time delays

This paper investigates the problem for stability of neutral-type dynamical neural networks involving delay parameters. Different form the previously reported results, the states of the neurons involve multiple delays and time derivative of states of neurons include discrete time delays. The stability of such neural systems has not been given much attention in the past literature due to the difficulty of finding Lyapunov functionals which are suitable for stability analysis of this type of neural networks. This paper constructs a generalized Lyapunov functional by introducing new terms into the well-known Lyapunov functional that enables us to conduct a theoretical investigation into stability analysis of delayed neutral-type neural systems. Based on this modified novel Lyapunov functional, sufficient criteria are derived, which guarantee the existence, uniqueness and global asymptotic stability of the equilibrium point of the neutral-type neural networks with multiple delays in the states and discrete delays in the time derivative of the states. The applicability of the proposed stability conditions rely on testing two basic matrix properties. The constraints impose on the system matrices are determined by using nonsingular M-matrix condition, and the constraints imposed on the coefficients of the time derivative of the delayed state variables are derived by exploiting the vector-matrix norms. We also note that the obtained stability conditions have no involvement with the delay parameters and expressed in terms of nonlinear Lipschitz activation functions. We present a constructive numerical example for this class of neural networks to give a systematic procedure for determining the imposed conditions on the whole system parameters of the delayed neutral-type neural systems.     

Automatic keyword identification by artificial neural networks compared to manual identification by users of filtering systems

Information filtering (IF) systems usually filter data items by correlating a vector of terms that represent the user profile with similar vectors of terms that represent data items. Terms that represent data items can be determined by experts or automatic indexing methods. In this study we employ an artificial neural network (ANN) as an alternative method for both IF and term selection and compare its effectiveness to that of “traditional” methods. In an earlier study we developed and examined the performance of an IF system that employed content-based and stereotypic rule-based filtering methods in the domain of e-mail messages. In this study, we train a large-scale ANN-based filter, which uses meaningful terms in the same database as input, and use it to predict the relevance of those messages. Our results reveal that the ANN relevance prediction out-performs the prediction of the IF system. Moreover, we found very low correlation between the terms in the user profile (explicitly selected by the users) and the positive causal-index (CI) terms of the ANN, which indicate the relative importance of terms in messages. This implies that the users underestimate the importance of some terms, failing to include them in their profiles. This may explain the rather low prediction accuracy of the IF system.     

神经网络方法在高原天气预报中的应用

本文应用了Rumelhart和McClelland于1985年发明的多层前馈网络算法（简称BP算法）。通过神经网络方法，对高原地区气温预报方面做了尝试，验证得出日极端最高气温预报比较理想，但是对日极端最低气温预报比前者要差一些。     

Listening to the investors: A novel framework for online lending default prediction using deep learning neural networks

Online peer-to-peer (P2P) lending has developed dramatically over the last decade in China. But this rapid boom carries potential risks. Investors have incurred incalculable losses due to the recent increase in fraudulent and/or unreliable online P2P platforms. Hence, predicting and identifying potential default risk platforms is crucial at this juncture. To achieve this end, we propose a two-step method which employs a deep learning neural network to extract keywords from investor comments and then utilizes a bidirectional long short-term memory (BiLSTM) based model to predict the default risk of platforms. Experimental results on real-world datasets of about 1000 platforms show that in the keyword extraction phase, our model can better capture semantic features from highly colloquial comment-text and achieve significant improvement over other baselines. Additionally, in the default platform prediction stage, our model achieves an F1 value of 80.34% in identifying potential problem platforms, outperforming four baselines by 23.37%, 5.71%, 8.93%, and 4.98% of improvement and comprehensively verifying the effectiveness of our method. Our study provides an alternative solution for platform default risk prediction issues and validates the effectiveness of investor comments in revealing the risk situation of online lending platforms.     

Stabilization effect of diffusion in delayed neural networks systems with Dirichlet boundary conditions

In this paper, reaction–diffusion neural networks with unbounded time-varying delays and Dirichlet boundary conditions is studied. A new concept of global μ-stability in the sense of L² norm is introduced, and sufficient conditions are given to guarantee global μ-stability of the equilibrium point. The results obtained not only improve those in the earlier findings, but also show diffusion terms contribute to stabilization of neural networks systems.     

Patterns of polarization: Transnational dynamics in climate change online networks in the US and Switzerland

Research into hyperlink interaction patterns has been particularly interested in whether they integrate the online space or segregate it into “echo chambers.” Concentrating on contentious politics in national settings, the existing studies have mainly examined the relationships between domestic actors, mostly bloggers. This study seeks to expand the focus by including several actor types, allowing their connective actions to reach beyond national borders, and employing a comparative approach that contrasts high- with low-contentious contexts. Analyzing climate change hyperlink networks originating in the US and Switzerland, the results show that their transnational dimension plays a crucial role in polarizing the discourse, regardless of the specific political context. We find similar patterns that segregate climate advocates from skeptics and lead to distinct transnational relationships within the camps. The results demonstrate that countermovement actors in particular are able to forge strong transnational alliances.     

Synchronization of fractional-order spatiotemporal complex-valued neural networks in finite-time interval and its application

This paper focuses on the synchronization of fractional-order complex-valued neural networks (FOCVNNs) with reaction–diffusion terms in finite-time interval. Different from the existing complex-valued neural networks (CVNNs), the reaction–diffusion phenomena and fractional derivative are first considered into the system, meanwhile, the parameter switching (the system parameters will switch with the state) is considered, which makes the presented model more comprehensive. By choosing an appropriate Lyapunov function, the driver and response systems achieve Mittag-Leffler synchronization under a suitable controller. In addition, based on the fractional calculus theorem and the basic inequality methods, a criterion of synchronization for the error system in finite-time interval is derived and the upper bound of the corresponding finite synchronization time can be obtained. Finally, two examples are provided, one is a numerical example to explain the effectiveness of the main results, and the other shows that the results of this paper can be applied to image encryption for any size with high-security coefficient.     

Multi-layer self-organizing polynomial neural networks and their development with the use of genetic algorithms

In this paper, we introduce a new architecture of genetic algorithms (GA)-based self-organizing polynomial neural networks (SOPNN) and discuss a comprehensive design methodology. Let us recall that the design of the “conventional” PNNs uses an extended group method of data handling (GMDH) and exploits polynomials (such as linear, quadratic, and modified quadratic functions) as well as considers a fixed number of input nodes (as being selected in advance by a network designer) at polynomial neurons (or nodes) located in each layer. The proposed GA-based SOPNN gives rise to a structurally optimized structure and comes with a substantial level of flexibility in comparison to the one we encounter in conventional PNNs. The design procedure applied in the construction of each layer of a PNN deals with its structural optimization involving the selection of preferred nodes (or PNs) with specific local characteristics (such as the number of input variables, the order of the polynomial, and a collection of the specific subset of input variables) and addresses specific aspects of parametric optimization. An aggregate performance index with a weighting factor is proposed in order to achieve a sound balance between approximation and generalization (predictive) abilities of the network. To evaluate the performance of the GA-based SOPNN, the model is experimented with using chaotic time series data. A comparative analysis reveals that the proposed GA-based SOPNN exhibits higher accuracy and superb predictive capability in comparison to some previous models available in the literature.     

基于神经网络的电离层F2层临界频率预测方法

在电离层风暴期,现存的电离层F2层临界频率预测方法不能满足实际应用的要求。根据磁层ap系数和太阳黑子月均值作为风暴期训练序列,本文提出了一种基于神经网络的电离层F2层临界频率预测新方法。模拟结果表明,这种新方法比现有的预测方法（STORM模型和Cander提出的神经网络方法）具有更好的预测性能。