Predicting cortical oscillations with bidirectional LSTM network: a simulation study

It has been stated that up-down-state (UDS) cortical oscillation levels between excitatory and inhibitory neurons play a fundamental role in brain network construction. Predicting the time series behaviors of neurons in periodic and chaotic regimes can help in improving diseases, higher-order human activities, and memory consolidation. Predicting the time series is usually done by machine learning methods. In paper, the deep bidirectional long short-term memory (DBLSTM) network is employed to predict the time evolution of regular, large-scale UDS oscillations produced by a previously developed neocortical network model. In noisy time-series prediction tasks, we compared the DBLSTM performance with two other variants of deep LSTM networks: standard LSTM, LSTM projected, and gated recurrent unit (GRU) cells. We also applied the classic seasonal autoregressive integrated moving average (SARIMA) time-series prediction method as an additional baseline. The results are justified through qualitative resemblance between the bifurcation diagrams of the actual and predicted outputs and quantitative error analyses of the network performance. The results of extensive simulations showed that the DBLSTM network provides accurate short and long-term predictions in both periodic and chaotic behavioral regimes and offers robust solutions in the presence of the corruption process.

     

Nested-spectral analysis reveals a disruption of behavioral-related dynamic functional balance in the aging brain

The relationship between the age-related reorganization of brain networks and individual behavior has attracted much attention. However, how age induces changes in neural activity at different frequencies in the brain to balance the demands of network integration and segregation, and how age-induced changes in network integration and segregation relate to behavior remain enigmatic. Here, a nested-spectral partition method was used to analyze behavioral-related dynamic functional balance in the aging brain with electroencephalogram signals collected from 56 healthy participants (age: 20–80 years) at rest. The nested-spectral partition approach measures hierarchical segregation and integration across multiple levels by detecting hierarchical modules in brain functional networks. Declines in general personality and general cognitive ability in older adults were captured by exploratory factor analysis. We showed that the brain network of elderly individuals contains more hierarchical modules to generate higher segregation, and it is closer to the functional balance state in the theta and alpha bands but away from this state in the gamma band. Meanwhile, the abnormal variability of functional balance in the elderly brain supports more flexible transitions between segregated and integrated states in the alpha band but reduces the transitions in the beta and gamma bands. Crucially, the degeneration of general personality and general cognitive ability is significantly associated with higher segregation and abnormal flexibility of the brain, especially in the theta, beta, and gamma bands. Our results provide deep insights from a spectral partitioning perspective into the brain dynamic mechanisms that are associated with age-related personality and cognitive degeneration.

     

Control and analysis of epilepsy waveforms in a disinhibition model of cortex network

Considering the disinhibition circuit between inhibitory neuronal populations with different time scales in cortical neural networks, here we propose a novel model to describe the occurrences and transitions of epilepsy waveforms. With the model we can successfully simulate poly-spike complexes, which are common in electrophysiological experiments and focal epilepsy patients. Meanwhile, we focus on the dynamic transitions between epilepsy waveforms and normal state and are devoted to exploring effective electrical stimulation strategies. Results show that disinhibition can induce an epileptic bidirectional transition, which is from spike and wave discharges, to poly-spike complexes and then to low-voltage rapid discharge activity, or it is reversed. And fascinating dynamical transition behaviors can be induced by varying average inhibitory synaptic gain. Interestingly, after applying two different control signals (deep brain stimulation and oscillatory input) to the system, all epilepsy waveforms can be suppressed or even eliminated. Results shed light on the pathophysiological mechanisms of epilepsy and guide clinical treatment from a theoretical viewpoint.

     

Hyperelastic model for polyacrylamide-gelatin double network shape-memory hydrogels

A shape-memory double network hydrogel consists of two polymer networks: a chemically crosslinked primary network that is responsible for the permanent shape and a physically crosslinked secondary network that is used to fix the temporary shapes. The formation/melting transition of the secondary network serves as an effective mechanism for the double network hydrogel's shape-memory effect. When the crosslinks in the secondary network are dissociated by applying an external stimulus, only the primary network is left to support the load. When the secondary network is re-formed by removing the stimulus, both the primary and secondary networks support the load. In the past, models have been developed for the constitutive behaviors of double network hydrogels, but the model of shape-memory double network hydrogels is still lacking. This work aims to build a constitutive model for the polyacrylamide-gelatin double network shape-memory hydrogel developed in our previous work. The model is first calibrated by experimental data of the double network shape-memory hydrogel under uniaxial loading and then employed to predict the shape-fixing performance of the hydrogel. The model is also implemented into a three-dimension finite element code and utilized to simulate the shape-memory behavior of the double network hydrogel with inhomogeneous deformations related to applications.

Graphic abstract

A shape-memory double network hydrogel consists of a chemically crosslinked primary network and a physically crosslinked secondary network. The formation/melting transition of the secondary network serves as an effective mechanism for the shape-memory effect of the double network hydrogel. This work built a constitutive model for the polyacrylamide-and-gelatin double network shape-memory hydrogel. The model was first calibrated by experimental data and then employed to predict the shape-fixing performance of the hydrogel. The model was also implemented into a three-dimension finite element code and utilized to simulate the shape-memory behavior of double network hydrogel in complex geometries.

     

Spread mechanism and control strategy of social network rumors under the influence of COVID-19

Since the outbreak of coronavirus disease in 2019 (COVID-19), the disease has rapidly spread to the world, and the cumulative number of cases is now more than 2.3 million. We aim to study the spread mechanism of rumors on social network platform during the spread of COVID-19 and consider education as a control measure of the spread of rumors. Firstly, a novel epidemic-like model is established to characterize the spread of rumor, which depends on the nonautonomous partial differential equation. Furthermore, the registration time of network users is abstracted as ‘age,’ and the spreading principle of rumors is described from two dimensions of age and time. Specifically, the susceptible users are divided into higher-educators class and lower-educators class, in which the higher-educators class will be immune to rumors with a higher probability and the lower-educators class is more likely to accept and spread the rumors. Secondly, the existence and uniqueness of the solution is discussed and the stability of steady-state solution of the model is obtained. Additionally, an interesting conclusion is that the education level of the crowd is an essential factor affecting the final scale of the spread of rumors. Finally, some control strategies are presented to effectively restrain the rumor propagation, and numerical simulations are carried out to verify the main theoretical results.

     

融合稀疏八叉树与卷积神经网络的汽车风阻系数预测

针对汽车风阻系数预测研究中参数化方法难以准确表征汽车外造型的难题,提出融合稀疏八叉树与卷积神经网络的汽车风阻系数预测方法。将汽车外造型按照八叉树结构离散,使用平均法向量对离散的复杂曲面进行简化,利用卷积神经网络对八叉树形式的汽车外造型进行特征提取,进而对汽车风阻系数进行快速预测。通过改变卷积层数与全连接层数,研究了不同卷积神经网络结构对风阻系数预测精度的影响。与参数化方法相比,本文提出的外造型表示方法能更好地描述模型细节,构建的卷积神经网络结构对风阻系数预测的最小相对误差为1.453%,且计算速度是CFD仿真的1620倍,具有较高的精度及计算效率。     

Structural reliability analysis for implicit performance function using radial basis function network

This is the second paper of our work on structural reliability analysis for implicit performance function. The first paper proposed structural reliability analysis methods using multilayer perceptron artificial neural network [Deng, J., Gu, D.S., Li, X.B., Yue, Z.Q., 2005. Structural reliability analysis for implicit performance function using artificial neural network. Structural Safety 25 (1), 25–48]. This paper presents three radial basis function network (RBF) based reliability analysis methods, i.e. RBF based MCS, RBF based FORM, and RBF based SORM. In these methods, radial basis function network technique is adopted to model and approximate the implicit performance functions or partial derivatives. The RBF technique uses a small set of the actual data of the implicit performance functions, which are obtained via physical experiments or normal numerical analysis such as finite element methods for the complicated structural system, and are used to develop a trained RBF generalization algorithm. Then a large number of the function values and partial derivatives of implicit performance functions can be readily obtained by simply extracting information from the established and successfully trained RBF network. These function values and derivatives are used in conventional MCS, FORM or SORM to constitute RBF based reliability analysis algorithms. Examples are presented in the paper to illustrate how the proposed RBF based methods are used in structural reliability analysis. The results are well compared with those obtained by the conventional reliability methods such as the Monte-Carlo simulation, multilayer perceptrons networks, the response surface method, the FORM method 2, and so on. The examples showed the proposed approach is applicable to structural reliability analysis involving implicit performance functions.     

The worst response of mistuned bladed disk system using neural network and genetic algorithm

The objective of this paper is to develop an integrated approach using artificial neural networks (ANN) and genetic algorithms (GA) for predicting the worst response of mistuned bladed disk. ANN is used to predict the responses of bladed disk system which are used further in evaluation of fitness and constraint violation in GA process. A multilayer back-propagation neural network is trained with the results obtained from finite element model for different bladed disk configurations. Subsequently, GA is employed for arriving at optimum configuration of the bladed disk system by maximizing the blade responses. By integrating ANN with GA, the computational time required for obtaining optimal solution could be reduced substantially. The efficacy of this approach is demonstrated by carrying out studies on mistuned bladed disk systems for different sets of mistuning parameters, namely mistuning in modulus of elasticity and length of blades. Finally, the effect of adding shroud at the tip of blades in reducing the maximum response of the bladed disk system was investigated.     

An adaptive vibrational resonance method based on cascaded varying stable-state nonlinear systems and its application in rotating machine fault detection

A weak character signal with low frequency can be detected based on the mechanism of vibrational resonance (VR). The detection performance of VR is determined by the synergy of a weak low-frequency input signal, an injected high-frequency sinusoidal interference and the nonlinear system(s). In engineering applications, there are many weak fault signals with high character frequencies. These fault signals are usually submerged in strong background noise. To detect these weak signals, an adaptive detection method for a weak high-frequency fault signal is proposed in this paper. This method is based on the mechanics of VR and cascaded varying stable-state nonlinear systems (VSSNSs). Partial background noise with high frequency is regarded as a special type of high-frequency interference and an energy source that protrudes a weak fault signal. In this way, high-frequency background noise is utilized in a VSSNS. To improve the detection ability, manually generated high-frequency interference is injected into another VSSNS. The VSSNS can be transformed into a monostable state, bistable state or tristable state by tuning the system parameters. The proposed method is validated by a simulation signal and industrial applications. The results show the effectiveness of the proposed method to detect a weak high-frequency character signal in engineering problems.

     

基于卷积神经网络的无网格形函数影响域优化研究

在无网格法中,离散节点之间的相互联系由节点形函数影响域的大小确定,因此形函数影响域的大小对无网格法的计算精度有着直接和重要的影响.但由于无网格形函数的形式较为复杂,目前形函数影响域大小的选择仍然缺乏系统的理论依据,通常在实际计算中仍凭借经验进行选取,难以保证计算精度.卷积神经网络是一类机器学习方法,其感受野与无网格形函...     

Second-order coupled tristable stochastic resonance and its application in bearing fault detection under different noises

Bearing fault is the most likely to occur in mechanical fault, and stochastic resonance (SR), as a noise enhanced signal processing tool, can find mechanical faults as early as possible, so as to avoid larger problems. However, most of the existing research methods are based on the first-order Langevin equation. According to the previous studies of many scholars, the weak signal detection ability of the second-order system is better than that of the first-order system, and the coupled system also has better performance due to the addition of the control system. So, in order to detect the fault signal more easily, a second-order coupled tristable stochastic resonance system (SCTSR) based on the adaptive genetic algorithm (AGA) is proposed, it is an improvement on improving the first-order coupled tristable stochastic resonance system (FCTSR). First, based on the fourth-order Runge–Kutta algorithm (F-RK), the performances of monostable, bistable and tristable control systems to SCTSR are compared, it is verified that the monostable system has the best performance as SCTSR’s control system. Secondly, the equivalent potential function of SCTSR is derived, and the influences of each system parameters on it are researched. The output signal-to-noise ratio gain (SNRG) is chosen as a measure to verify that SCTSR’s performance is better than that of FCTSR, and the influences of parameters on SNRG are discussed. SCTSR and FCTSR are used to detect low-, high- and multi-frequency cosine signals combined with AGA. The simulation results are compared with the wavelet transform method, which proves the performance superiority of SR, and also prove that SCTSR is easier to detect weak signals and has a stronger de-noising ability. Finally, SCTSR and FCTSR are applied in bearing fault detection under Gaussian white noise and trichotomous noise. The results also prove that SCTSR can get larger peaks and SNRG, and it is easier to detect fault signals. This proves that SCTSR’s performance is superior that of other methods in bearing fault detection, and has better engineering application value.

     

A simulation of the cascading failure of a complex network model by considering the characteristics of road traffic conditions

Using the dual method, we start with a traditional road traffic network with a constructed logic network with small-world characteristics and construct the complex network of road traffic. After analyzing and comparing with other complex networks, the time delay, restorative, and other characteristics are presented for the complex network of road traffic, and then, the cascading failure model of the complex network is simulated. The simulation results show that using different time delays, an incident dissipation factor and load capacity can reasonably avoid a cascading failure, and they can remove its effects. In addition, our results provide value and guidance for building a road traffic network that prevents and removes the cascading failure of a road network.

     

Gas–liquid flows in a microscale fractal-like branching flow network

Two-phase air–water flows in a microscale fractal-like flow network were experimentally studied and results were compared to predictions from existing macroscale void fraction correlations and flow regime maps. Void fraction was assessed using (1) two-dimensional analysis of high-speed images (direct method) and (2) experimentally determined using gas velocities (indirect method). Fixed downstream-to-upstream length and width ratios of 1.4 and 0.71, respectively, characterize the five-level flow network. Channels were fabricated in a 38 mm diameter silicon disk, 250 μm deep disk with a terminal channel width of 100 μm. A Pyrex top allowed for flow visualization. Superficial air and water velocities through the various branch levels were varied from 0.007 m/s to 1.8 m/s and from 0.05 m/s to 0.42 m/s, respectively. Two-phase flow regime maps were generated for each level of the flow network and are well predicted by the Taitel and Dukler model. Void fraction assessed using the indirect method shows very good agreement with the homogeneous void fraction model for all branch levels for the given range of flow conditions. Void fraction determined directly varies considerably from that assessed indirectly, showing better agreement with the void fraction correlation of Zivi.     

Memristive hyperchaos secure communication based on sliding mode control

Nonlinear Dynamics - In order to enhance the chaotic degree of cellular neural network (CNN), the memristive characteristic is combined in CNN, and a five-dimensional memristive CNN hyperchaotic...     

Relaxation and Recovery in Hydrogel Friction on Smooth Surfaces

Background

Hydrogels are crosslinked polymer networks that can absorb and retain a large fraction of liquid. Near a critical sliding velocity, hydrogels pressed against smooth surfaces exhibit time-dependent frictional behavior occurring over multiple timescales. The origin of these dynamics is unresolved

Objective

Here, we characterize this time-dependent regime and show that it is consistent with two distinct molecular processes: sliding-induced relaxation and quiescent recovery.

Methods

Our experiments use a custom pin-on-disk tribometer to examine poly(acrylic acid) hydrogels on smooth poly(methyl methacrylate) surfaces over a variety of sliding conditions, from minutes to hours.

Results

We show that at a fixed sliding velocity, the friction coefficient decays exponentially and reaches a steady-state value. The time constant associated with this decay varies exponentially with the sliding velocity, and is sensitive to any precedent frictional shearing of the interface. This process is reversible; upon cessation of sliding, the friction coefficient recovers to its original state. We also show that the initial direction of shear can be imprinted as an observable “memory”, and is visible after 24 hrs of repeated frictional shearing.

Conclusions

We attribute this behavior to nanoscale extension and relaxation dynamics of the near-surface polymer network, leading to a model of frictional relaxation and recovery with two parallel timescales.

     

Data-driven solutions and parameter discovery of the nonlocal mKdV equation via deep learning method

In this paper, we systematically study the integrability and data-driven solutions of the nonlocal mKdV equation. The infinite conservation laws of the nonlocal mKdV equation and the corresponding infinite conservation quantities are given through Riccti equation. The data-driven solutions of the zero boundary for the nonlocal mKdV equation are studied by using the multilayer physical information neural network algorithm, which include kink soliton, complex soliton, bright-bright soliton and the interaction between soliton and kink-type. For the data-driven solutions with nonzero boundary, we study kink, dark, anti-dark and rational solution. By means of image simulation, the relevant dynamic behavior and error analysis of these solutions are given. In addition, we discuss the inverse problem of the integrable nonlocal mKdV equation by applying the physics-informed neural network algorithm to discover the parameters of the nonlinear terms of the equation.

     

Physics-constrained bayesian neural network for fluid flow reconstruction with sparse and noisy data

In many applications, flow measurements are usually sparse and possibly noisy. The reconstruction of a high-resolution flow field from limited and imperfect flow information is significant yet challenging. In this work, we propose an innovative physics-constrained Bayesian deep learning approach to reconstruct flow fields from sparse, noisy velocity data, where equationbased constraints are imposed through the likelihood function and uncertainty of the reconstructed flow can be estimated. Specifically, a Bayesian deep neural network is trained on sparse measurement data to capture the flow field. In the meantime, the violation of physical laws will be penalized on a large number of spatiotemporal points where measurements are not available. A non-parametric variational inference approach is applied to enable efficient physicsconstrained Bayesian learning. Several test cases on idealized vascular flows with synthetic measurement data are studied to demonstrate the merit of the proposed method.     

Fast Flow Field Estimation for Various Applications with A Universally Applicable Machine Learning Concept

This paper presents an approach for the prediction of incompressible laminar steady flow fields over various geometry types. In conventional approaches of computational fluid dynamics (CFD), flow fields are obtained by solving model equations on computational grids, which is in general computationally expensive. Based on the ability of neural networks to intuitively identify and approximate nonlinear physical relationships, the proposed method makes it possible to eliminate the explicit implementation of model equations such as the Navier–Stokes equations. Moreover, it operates without iteration or spatial discretization of the flow problem. The method is based on the combination of a minimalistic multilayer perceptron (MLP) architecture and a radial-logarithmic filter mask (RLF). The RLF acts as a preprocessing step and its purpose is the spatial encoding of the flow guiding geometry into a compressed form, that can be effectively interpreted by the MLP. The concept is applied on internal flows as well as on external flows (e.g. airfoils and car shapes). In the first step, datasets of flow fields are generated using a CFD-code. Subsequently the neural networks are trained on defined portions of these datasets. Finally, the trained neural networks are applied on the remaining unknown geometries and the prediction accuracy is evaluated. Dataset generation, neural network implementation and evaluation are carried out in MATLAB. To ensure reproducibility of the results presented here, the trained neural networks and sample applications are made available for free download and testing.

     

The complete synchronization condition in a network of piezoelectric micro-beams

This work deals with the dynamics of a network of piezoelectric micro-beams (a stack of disks). The complete synchronization condition for this class of chaotic nonlinear electromechanical system with nearest-neighbor diffusive coupling is studied. The nonlinearities within the devices studied here are in both the electrical and mechanical components. The investigation is made for the case of a large number of coupled discrete piezoelectric disks. The problem of chaos synchronization is described and converted into the analysis of the stability of the system via its differential equations. We show that the complete synchronization of N identical coupled nonlinear chaotic systems having shift invariant coupling schemes can be calculated from the synchronization of two of them. According to analytical, semi-analytical predictions and numerical calculations, the transition boundaries for chaos synchronization state in the coupled system are determined as a function of the increasing number of oscillators.

     

Solving Huxley equation using an improved PINN method

Although many effective methods for solving partial differential equations (PDEs) have been proposed, there is no universal method that can solve all PDEs. Therefore, solving partial differential equations has always been a difficult problem in mathematics, such as deep neural network (DNN). In recent years, a method of embedding some basic physical laws into traditional neural networks has been proposed to reveal the dynamic behavior of equations directly from space-time data [i.e., physics-informed neural network (PINN)]. Based on the above, an improved deep learning method to recover the new soliton solution of Huxley equation has been proposed in this paper. As far as we know, this is the first time that we have used an improved method to study the numerical solution of the Huxley equation. In order to illustrate the advantages of the improved method, we use the same network depth, the same hidden layer and neurons contained in the hidden layer, and the same training sample points. We analyze the dynamic behavior and error of Huxley’s exact solution and the new soliton solution and give vivid graphs and detailed analysis. Numerical results show that the improved algorithm can use fewer sample points to reconstruct the exact solution of the Huxley equation with faster convergence speed and better simulation effect.