Experimental verification of a memristive neural network

This paper presents an electronic circuit able to emulate the behavior of a neural network based on memristive synapses. The latter is built with two flux-controlled floating memristor emulator circuits operating at high frequency and two passive resistors. Synapses are connected in a way that a bridge circuit is obtained, and its dynamical behavioral model is derived from characterizing memristive synapses. Analysis of the memristor characteristics for obtaining a suitable synaptic response is also described. A neural network of one neuron and two inputs is connected using the proposed topology, where synaptic positive and negative weights can easily be reconfigured. The behavior of the proposed artificial neural network based on memristors is verified through MATLAB, HSPICE simulations and experimental results. Synaptic multiplication is performed with positive and negative weights, and its behavior is also demonstrated through experimental results getting 6% of error.     

Chimera states and synchronization behavior in multilayer memristive neural networks

To study the effect of electromagnetic induction on the spatiotemporal dynamic behavior of neural networks, in this paper, we have mainly studied both the synchronization behavior and the evolution of chimera states (CS) in coupled neural networks. To do this, a multilayer memristive neural network was constructed by selecting the Hindmarsh–Rose neurons as the network nodes, and the effect of electromagnetic induction is introduced by using the cubic flux-controlled memristive model as synapse. For simplicity, the following coupling scheme is adopted: only the coupling connections for the neurons between different layers are considered with memristive synapses, while those neurons in each layer are still bidirectional coupled with the classical electrical synapses. It is found that, by adjusting the coupled strength of electrical synapses and the parameters of memristive synapses, the coexistence behavior of coherent and incoherent states, i.e., the CS, could appear in each layer. It is interesting that the CS are also found in inter-layer memristive synapse network. Furthermore, we have discussed the synchronization behavior in this multilayer memristive neural network, one can find when the whole multilayer network is in a synchronization state, not only the spatiotemporal consistency of the CS in each layer neural networks is observed, but also the memductance of all memristive synapses is completely synchronized. Our results suggest that the electromagnetic induction may play an important role in regulating the dynamic behavior of neural networks, and the introduction of memristive synapse into the biological neural network will provide useful clues for revealing the memory behavior of the neural system in human brain.     

The collective bursting dynamics in a modular neuronal network with synaptic plasticity

In this study, how the synaptic plasticity influences the collective bursting dynamics in a modular neuronal network is numerically investigated. The synaptic plasticity is described by a modified Oja’s learning rule. The modular network is composed of some sub-networks, each of them having small-world characteristic. The result indicates that bursting synchronization can be induced by large coupling strength between different neurons, which is robust to the local dynamical parameter of individual neurons. With the emergence of synaptic plasticity, the bursting dynamics in the modular neuronal network, particularly the excitability and synchronizability of bursting neurons, is detected to be changed significantly. In detail, upon increasing synaptic learning rate, the excitability of bursting neurons is greatly enhanced; on the contrary, bursting synchronization between interacted neurons is a little suppressed by the increase in synaptic learning rate. The presented findings could be helpful to understand the important role of synaptic plasticity on neural coding in realistic neuronal network.     

结构试验机的神经网络内模自适应跟踪控制系统

针对结构试验系统的非线性和不确定性特性，研究了一种基于神经网络的非线性内模自适应加载控制方法。引入的神经网络内模可跟踪学习对象的时变动力学，控制器的设计较少依赖于对象的先验知识，控制参数的调整是基于被控过程的测量信息，利用导出的神经网络算法来实现的。实验结果证明该系统具有良好的控制效果。     

运用机器学习方法设计原子链 人工边界条件

本文运用机器学习方法设计一维线性原子链的人工边界条件。该方法基于前馈神经网络,通过对一小部分数值解进行训练后得到人工边界条件。应用该法不需要较多先验知识,编程简单,实现速度快,算例表明数值反射较小。     

基于卷积神经网络的涵洞式直立堤波浪透射预测

涵洞式直立堤是一种具有特殊用途的海岸工程结构物,对其透浪特性的研究具有重要工程意义.然而,目前众多学者对于涵洞式直立堤波浪透射问题的研究主要以理论分析、实验模拟及数值计算为主.随着机器学习技术的发展,传统水动力学问题迎来了新的求解理念.机器学习算法可根据训练数据集自主学习相应的规律,以数据映射的方式建立水动力学特征预测...     

Physics-informed deep learning for digital materials

In this work, a physics-informed neural network(PINN) designed specifically for analyzing digital materials is introduced. This proposed machine learning(ML) model can be trained free of ground truth data by adopting the minimum energy criteria as its loss function. Results show that our energy-based PINN reaches similar accuracy as supervised ML models. Adding a hinge loss on the Jacobian can constrain the model to avoid erroneous deformation gradient caused by the nonlinear logarithmic strain. Lastly, we discuss how the strain energy of each material element at each numerical integration point can be calculated parallelly on a GPU. The algorithm is tested on different mesh densities to evaluate its computational efficiency which scales linearly with respect to the number of nodes in the system. This work provides a foundation for encoding physical behaviors of digital materials directly into neural networks, enabling label-free learning for the design of next-generation composites.     

Using a Lumped Parameter Dynamic Model of a Rock Bolt to Produce Training Data for a Neural Network for Diagnosis of Real Data

Ground anchorage systems are used extensively throughout the world as supporting devices for civil engineering structures such as bridges and tunnels. The condition monitoring of ground anchorages is a new area of research, with the long term objective being a wholly automated or semi-automated condition monitoring system capable of repeatable and accurate diagnosis of faults and anchorage post-tension levels. The ground anchorage integrity testing (GRANIT) system operates by applying an impulse of known force by means of an impact device that is attached to the tendon of the anchorage. The vibration signals that arise from this impulse are complex in nature and require analysis to be undertaken in order to extract information from the vibrational response signatures that is relevant to the condition of the anchorage. Novel artificial intelligence techniques are used in order to learn the complicated relationship that exists between an anchorage and its response to an impulse. The system has a worldwide patent and is currently licensed commercially.A lumped parameter dynamic model has been developed which is capable of describing the general frequency relationship with increasing post-tension level as exhibited by the signals captured from real anchorages. The normal procedure with the system is to train a neural network on data that has been taken from an anchorage over a range of post-tension levels. Further data is needed in order to test the neural network. This process can be time consuming, and the lumped parameter dynamic model has the potential of producing data that could be used for training purposes, thereby reducing the amount of time needed on site, and reducing the overall cost of the system's operation.This paper presents data that has been produced by the lumped parameter dynamic model and compares it with data from a real anchorage. Noise is added to the results produced by the lumped parameter dynamic model in order to match more closely the experimental data. A neural network is trained on the data produced by the model, and the results of diagnosis of real data are presented. Problems are encountered with the diagnosis of the neural network with experimental data, and a new method for the training of the neural network is explored. The improved results of the neural network trained on data produced by the lumped parameter dynamic model to experimental data are shown. It is shown how the results from the lumped parameter dynamic model correspond well to the experimental results.     

边坡稳定性预测的模糊神经网络模型

根据边坡稳定问题具有的模糊性,提出了一种判定边坡稳定性的模糊神经网络模型。该系统仅从期望输入输出数据集即可达到获取知识、确定模糊初始规则基的目的。再利用神经网络学习能力便不难修改规则库中的模糊规则以及隶属函数和网络权值等参数,这样大大减少了规则匹配过程,加快了推理速度,从而极大程度地提高了系统的自适应能力。最后用收集到的边坡数据样本训练和测试模糊神经网络模型,结果表明该模糊神经网络预测边坡稳定性是可行的、有效的。     

基于人工神经网络的湍流大涡模拟方法

大涡模拟方法(LES)是研究复杂湍流问题的重要工具,在航空航天、湍流燃烧、气动声学、大气边界层等众多工程领域中具有广泛的应用前景.大涡模拟方法采用粗网格计算大尺度上的湍流结构,并用亚格子(SGS)模型近似表达滤波尺度以下的流动结构对大尺度流场的作用.传统的亚格子模型由于只利用了单点流场信息和简单的函数关系,在先验验证中相对误差较大, 在后验验证中耗散过强. 近几年来,机器学习方法在湍流建模问题中得到了越来越多的应用.本文介绍了基于人工神经网络(ANN)的湍流亚格子模型的最新进展.详细地讨论了人工神经网络混合模型、空间人工神经网络模型和反卷积人工神经网络模型的构造方法.借助于人工神经网络强大的数据插值能力,新的亚格子模型的先验精度和后验精度均有显著提升. 在先验验证中,新模型所预测的亚格子应力的相关系数超过了0.99,在预测精度上远高于传统的大涡模拟模型. 在后验验证中,新模型对各类湍流统计量和瞬态流动结构的预测都优于隐式大涡模拟方法、动态Smagorinsky模型、动态混合模型等传统模型.因此, 人工神经网络方法在发展复杂湍流的先进大涡模拟模型中具有很大的潜力.      

基于主动学习神经网络的转向架构架疲劳可靠性分析

为了提高转向架构架疲劳可靠性分析的精度与效率,提出一种主动学习BR-BP神经网络模型与Monte Carlo法相结合的可靠性分析方法。该方法针对BP神经网络的缺陷,使用贝叶斯正则BR(Bayesian regularization)算法作为训练算法,以提高神经网络的拟合精度与收敛速度,并考虑可靠性分析的固有特点,构造了一种适用于BP神经网络的主动学习函数,用于指导最佳样本点的选择。提出的学习函数不仅保证了样本点分布在极限状态函数附近,还考虑了样本点的预测误差以及样本点分布对失效概率计算的影响。转向架构架可靠性分析结果表明,本文方法在提高拟合精度的同时兼顾了计算效率,验证了所提方法的优越性与可行性。     

Predicting solutions of the Lotka‐Volterra equation using hybrid deep network

Prediction of Lotka-Volterra equations has always been a complex problem due to their dynamic properties. In this paper, we present an algorithm for predicting the Lotka-Volterra equation and investigate the prediction for both the original system and the system driven by noise. This demonstrates that deep learning can be applied in dynamics of population. This is the first study that uses deep learning algorithms to predict Lotka-Volterra equations. Several numerical examples are presented to illustrate the performances of the proposed algorithm, including Predator nonlinear breeding and prey competition systems, one prey and two predator competition systems, and their respective systems. All the results suggest that the proposed algorithm is feasible and effective for predicting Lotka-Volterra equations. Furthermore, the influence of the optimizer on the algorithm is discussed in detail. These results indicate that the performance of the machine learning technique can be improved by constructing the neural networks appropriately.     

On hyperchaos in a small memristive neural network

This paper studies a small Hopfield neural network with a memristive synaptic weight. We show that the previous stable network after one weight replaced by a memristor can exhibit rich complex dynamics, such as quasi-periodic orbits, chaos, and hyperchaos, which suggests that the memristor is crucial to the behaviors of neural networks and may play a significant role. We also prove the existence of a saddle periodic orbit, and then present computer-assisted verification of hyperchaos through a homoclinic intersection of the stable and unstable manifolds, which gives a positive answer to an interesting question that whether a 4D memristive system with a line of equilibria can demonstrate hyperchaos.     

无网格动力分析的循环卷积神经网络代理模型

在无网格动力分析中,除了无网格形函数本身构造复杂引入的计算成本,还需要逐步递推求解每个时间步的动力响应,因而计算效率较为低下.本文通过研究无网格离散数据与机器学习训练样本、无网格动力分析递推计算过程与循环卷积神经网络序列信息传递模式之间的本征联系,构建了与无网格法相匹配的循环卷积神经网络设计方法,进而提出了一种无网格动...     

神经网络时滞系统非共振双Hopf分岔及其广义同步

本文建立了具有自连接和抑制-兴奋型他连接的两个同性神经元模型。其中自连接是由于兴奋型的突触产生,而他连接则分别对应于两神经元兴奋、抑制型的突触。发现如果有兴奋型自连接就会有双Hopf分岔,而没有时滞自连接时双Hopf分岔就会消失,因此自连接引起了双Hopf分岔。作为一个例子,通过变动连接中的时滞和他连接中的比重,1／√2双Hopf分岔得到了详细研究。通过中心流形约化,分岔点邻域内各种不同的动力学行为得到了分类,并以解析形式表出。神经元活动的分岔路径得以表明。从得到的解析近似解可以发现,本文所研究的具有兴奋一抑制型他连接的两相同神经元的节律不能完全同步而只能广义同步。时滞也可以使其节律消失,两神经元变为非活动的。这些结果在控制神经网络关联记忆和设计人工神经网络方面有着潜在的应用。     

CNN/Kalman技术在组导中的应用研究

经典的卡尔曼滤波器要求假设系统的动态模型和观测模型的噪声统计特性已知，而组合导航系统的噪声具有非先验性。为了解决这一问题，提出了一种新型复合神经网络(CNN)辅助卡尔曼滤波器(Kalman)。仿真试验结果表明：该辅助算法的精度与一般卡尔曼算法相比提高了2倍，收敛时间缩短近200s，并有效地克服了传统神经网络学习速度慢、泛化能力弱的缺点，使系统具有自适应能力以应付动态环境的扰动。     

Condition Monitoring of Ground Anchorages by Dynamic Impulses: GRANIT System

The GRANIT system operates by applying an impulse of known force by means of an impact device that is attached to the tendon of the anchorage. The vibration response signals resulting from this impulse are complex in nature and require analysis to be undertaken in order to extract information from the vibrational response signatures that is relevant to the condition of the anchorage. In the system, the complicated relationship that exists between characteristics of an anchorage and its response to an impulse is identified and learned by a novel artificial intelligence network based on artificial intelligence techniques.The results presented in this paper demonstrate the potential of the GRANIT system to diagnose the integrity of ground anchorages at a site near Stone, England, by using a trained neural network capable of diagnosing the post-tension level of the anchorage. This neural network was used for the diagnosis of load in a second ground anchorage adjacent to the original anchorage used for the training of the neural network. Further tests were taken with a different anchor head configuration of the anchorage and a different relationship between the signature response of the anchorage to an applied impulse and its post-tension level was found.Problems encountered during the diagnosis of this second set of test signatures by the trained neural network are investigated with the use of a lumped parameter dynamic model. This model is able to identify the parameters in the anchorage system that affect this change in response signature. The results from the investigation lead to a new form of classification for the installed anchorages, based on their anchor head configuration.Laboratory strand anchorage tests were undertaken in order to compare with and validate the results obtained from the field tests and the lumped parameter dynamic model.     

Structure and dynamics of self-organized neuronal network with an improved STDP rule

Nonlinear Dynamics - The chemical synapses in a neural network are known to be modulated by the neuronal firing activities through the spike-timing-dependent plasticity (STDP) rule. In this paper,...     

Neural network ℋ<Subscript>∞</Subscript> chaos synchronization

This paper proposes a new neural network ?_∞ synchronization (NNHS) scheme for unknown chaotic systems. In the proposed framework, a dynamic neural network is constructed as an alternative to approximate the chaotic system. Based on this neural network and linear matrix inequality (LMI) formulation, the NNHS controller and the learning law are presented to reduce the effect of disturbance to an ?_∞ norm constraint. It is shown that finding the NNHS controller and the learning law can be transformed into the LMI problem and solved using the convex optimization method. A numerical example is presented to demonstrate the validity of the proposed NNHS scheme.     

基于神经网络的偏微分方程求解方法研究综述

神经网络作为一种强大的信息处理工具在计算机视觉,生物医学,油气工程领域得到广泛应用,引发多领域技术变革.深度学习网络具有非常强的学习能力,不仅能发现物理规律,还能求解偏微分方程.近年来基于深度学习的偏微分方程求解已是研究新热点.遵循于传统偏微分方程解析解、偏微分方程数值解术语,本文称用神经网络进行偏微分方程求解的方法为...