Nonlinear dynamics extraction for time-delay systems using modular neural networks synchronization and prediction

It is shown that the nonlinear dynamics of chaotic time-delay systems can be reconstructed using a new type of neural network with two modules: one for nonfeedback part with input data delayed by the embedding time, and a second one for the feedback part with input data delayed by the feedback time. The method is applied to both simulated and experimental data from an electronic analog circuit of the Mackey–Glass system. Better results are obtained for the modular than for feedforward neural networks for the same number of parameters. It is found that the complexity of the neural network model required to reconstruct nonlinear dynamics does not increase with the delay time. Synchronization between the data and the model with diffusive coupling is also achieved. We have also shown by iterating the model from the present point that the dynamics can be predicted with a forecast horizon larger than the feedback delay time.     

托卡马克中由线积分数据重建分布的神经网络方法

用多层前馈神经网络,处理托卡马克中由物理量沿观察弦的线积分值重建其空间分布的反演问题。用BFGS拟牛顿法大大加快了标准误差反向传播算法(BP)的收敛速度     

DeepcomplexMRI: Exploiting deep residual network for fast parallel MR imaging with complex convolution

This paper proposes a multi-channel image reconstruction method, named DeepcomplexMRI, to accelerate parallel MR imaging with residual complex convolutional neural network. Different from most existing works which rely on the utilization of the coil sensitivities or prior information of predefined transforms, DeepcomplexMRI takes advantage of the availability of a large number of existing multi-channel groudtruth images and uses them as target data to train the deep residual convolutional neural network offline. In particular, a complex convolutional network is proposed to take into account the correlation between the real and imaginary parts of MR images. In addition, the k-space data consistency is further enforced repeatedly in between layers of the network. The evaluations on in vivo datasets show that the proposed method has the capability to recover the desired multi-channel images. Its comparison with state-of-the-art methods also demonstrates that the proposed method can reconstruct the desired MR images more accurately.     

Phase Space Prediction Model Using Neural Networks

A new nonlinear prediction technique is proposed by feedforward neural network, the learning algorithmfor network is a chaotic one. A timc-delay embedding is used to reconstruct the underlying attractor, the predictionmodel is based on the time evolution of the topological neighboring in the phase space, the spatial neighbors are chosenby the rate of exponential divergence of close trajectory. The model is tested for the Mackey-Glass delay equation andLorentz equations, good results are obtained for the prediction.     

Bidirectional Information Flow Quantum State Tomography

The exact reconstruction of many-body quantum systems is one of the major challenges in modern physics,because it is impractical to overcome the exponential complexity problem brought by high-dimensional quantum manybody systems.Recently,machine learning techniques are well used to promote quantum information research and quantum state tomography has also been developed by neural network generative models.We propose a quantum state tomography method,which is based on a bidirectional gated recurrent unit neural network,to learn and reconstruct both easy quantum states and hard quantum states in this study.We are able to use fewer measurement samples in our method to reconstruct these quantum states and to obtain high fidelity.     

基于深度神经网络的时空编码磁共振成像超分辨率重建方法

单扫描时空编码磁共振成像是一种新型超快速磁共振成像技术,它对磁场不均匀和化学位移伪影有较强的抵抗性,但是其固有的空间分辨率较低,因此通常需要进行超分辨率重建,以在不增加采样点数的情况下提高时空编码磁共振图像的空间分辨率.然而,现有的重建方法存在迭代求解时间长、重建结果有混叠伪影残留等问题.为此,本文提出了一种基于深度神经网络的单扫描时空编码磁共振成像超分辨率重建方法.该方法采用模拟样本训练深度神经网络,再利用训练好的网络模型对实际采样信号进行重建.数值模拟、水模和活体鼠脑的实验结果表明,该方法能快速重建出无残留混叠伪影、纹理信息清楚的超分辨率时空编码磁共振图像.适当增加训练样本数量以及在训练样本中加入适当的随机噪声水平,有助于改善重建效果.     

A Nonlinear Maximum Correntropy Information Filter for High-Dimensional Neural Decoding

Neural signal decoding is a critical technology in brain machine interface (BMI) to interpret movement intention from multi-neural activity collected from paralyzed patients. As a commonly-used decoding algorithm, the Kalman filter is often applied to derive the movement states from high-dimensional neural firing observation. However, its performance is limited and less effective for noisy nonlinear neural systems with high-dimensional measurements. In this paper, we propose a nonlinear maximum correntropy information filter, aiming at better state estimation in the filtering process for a noisy high-dimensional measurement system. We reconstruct the measurement model between the high-dimensional measurements and low-dimensional states using the neural network, and derive the state estimation using the correntropy criterion to cope with the non-Gaussian noise and eliminate large initial uncertainty. Moreover, analyses of convergence and robustness are given. The effectiveness of the proposed algorithm is evaluated by applying it on multiple segments of neural spiking data from two rats to interpret the movement states when the subjects perform a two-lever discrimination task. Our results demonstrate better and more robust state estimation performance when compared with other filters.     

A deep learning–based method for improving reliability of multicenter diffusion kurtosis imaging with varied acquisition protocols

Multicenter magnetic resonance imaging is gaining more popularity in large-sample projects. Since both varying hardware and software across different centers cause unavoidable data heterogeneity across centers, its impact on reliability in study outcomes has also drawn much attention recently. One fundamental issue arises in how to derive model parameters reliably from image data of varying quality. This issue is even more challenging for advanced diffusion methods such as diffusion kurtosis imaging (DKI). Recently, deep learning–based methods have been demonstrated with their potential for robust and efficient computation of diffusion-derived measures. Inspired by these approaches, the current study specifically designed a framework based on a three-dimensional hierarchical convolutional neural network, to jointly reconstruct and harmonize DKI measures from multicenter acquisition to reformulate these to a state-of-the-art hardware using data from traveling subjects. The results from the harmonized data acquired with different protocols show that: 1) the inter-scanner variation of DKI measures within white matter was reduced by 51.5% in mean kurtosis, 65.9% in axial kurtosis, 53.7% in radial kurtosis, and 61.5% in kurtosis fractional anisotropy, respectively; 2) data reliability of each single scanner was enhanced and brought to the level of the reference scanner; and 3) the harmonization network was able to reconstruct reliable DKI values from high data variability. Overall the results demonstrate the feasibility of the proposed deep learning–based method for DKI harmonization and help to simplify the protocol setup procedure for multicenter scanners with different hardware and software configurations.     

混沌光学系统之前向神经网络混沌加速的系统辨识研究

研究了利用前向神经网络对混沌光学系统进行混沌加速系统辨识的可能性，计算机数值仿真发现，利用三层前向神经网络混沌光学系统辨识器。在基于混沌动力学角度的修正ＢＰ算法（混沌加速ＢＰ算法）支持下可克服由常规ＢＰ算法导致的辨识时间长的缺点，在较少的训练次数内即可对布拉格声双稳混沌系统进行良好的系统辨识，此研究结果表明，在混沌加速ＢＰ算法支持下，三层前向神经网络可用来快速处理混沌光学时间序列以进行相应的动力学     

多局部残差连接注意网络的图像去模糊

针对现有的基于卷积神经网络的图像去模糊算法存在图像纹理细节恢复不清晰的问题,提出了一种基于多局部残差连接注意网络的图像去模糊算法。首先,采用一个卷积层进行浅层特征提取;其次,设计了一种新的基于残差连接和并行注意机制的多局部残差连接注意模块,用于消除图像模糊并提取上下文信息;再次,采用一个基于扩张卷积的成对连接模块进行细节恢复;最后,利用一个卷积层重建清晰图像。实验结果表明：在GoPro数据集上的PSNR (peak signal to noise ratio)和SSIM (structure similarity)分别为31.83 dB、0.927 5,在定性和定量两方面都表明所提方法能够有效地恢复模糊图像的纹理细节,网络性能优于对比方法。     

Knowledge-Based Artificial Neural Network Models for Finline

Finline plays an important role in millimeter-wave integrated-circuit design. In this paper, a knowledge-based artifcial neural network is used to model the finline. Using prior knowledge input method and Bayesian regularization technique make the neural network models for finline reduce the amount of training data needed and prevent overfitting in neural network training. The neural network is electromagnetically developed with a set of training data that are produced by the fnite element method, which is robust both from the angle of time of computation and accuracy.     

Neural network applied to reconstruction of complex objects based on fringe projection

The neural network has been introduced into the reconstruction of the complex object based on fringe projection. In this method, the neural network with powerful property of approximation is used to get the continuous approximate function of a discrete fringe pattern captured by an image frame grabber. The depth-related phase of the measured object modulated into the fringe pattern can be demodulated by dealing with the approximate function. Compared with the Fourier transform profilometry (FTP), in the network method, one deformed fringe pattern is needed to reconstruct the tested object, and a high spatial resolution is maintained for no filtering process. Therefore, this method performs better than FTP in the measurement of the complex object. Moreover, the network method is capable of demodulating more depth-related phase even in the case that the local shadow exists in the fringe pattern. Computer simulations and experiments validate the feasibility of this method.     

模糊对向传播网络数据融合目标识别方法

重点研究了模糊对向传播网络 (FCPN)模型。针对数据融合和目标识别的特点 ,提出了基于模糊对向传播网络的融合目标识别方法和改进的模糊对向传播网络 (MFCPN)融合结构。利用仿真数据对网络的训练算法和融合结构进行了实验研究。结果表明 ,模糊对向传播网络较误差后向传播网络 (BPN)能够有效地实现融合识别 ;改进的模糊对向传播网络融合结构是可行的。同时还对FCPN和MFCPN应用于前视红外 (FLIR)和可见光摄像机目标跟踪系统进行了应用研究。     

Input reconstruction of chaos sensors

Although the sensitivity of sensors can be significantly enhanced using chaotic dynamics due to its extremely sensitive dependence on initial conditions and parameters, how to reconstruct the measured signal from the distorted sensor response becomes challenging. In this paper we suggest an effective method to reconstruct the measured signal from the distorted (chaotic) response of chaos sensors. This measurement signal reconstruction method applies the neural network techniques for system structure identification and therefore does not require the precise information of the sensor's dynamics. We discuss also how to improve the robustness of reconstruction. Some examples are presented to illustrate the measurement signal reconstruction method suggested.     

应用支持向量机和人工神经网络对大气次声信号 识别的初步实验

语音情感识别在许多领域具有重要研究价值,不同声学情感特征在使用不同分类器进行分类时,识别效果具有明显差异。与语音情感有关的声学特征包括谱特征、韵律学特征、音质特征。该文提出一种特征融合的方法,将3种声学特征中具有最好识别能力的特征进行融合:保留在实验中表现稳定且有较高识别率的谱特征的全部特征,提取韵律学、音质特征的相关统计量作为辅助特征融合于谱特征中。实验表明,该文所提出的融合特征在使用同一分类器进行分类时,识别率优于单一特征;当使用不同分类器时,融合特征依然具有较好的识别能力,且识别性能稳定,3个数据集上均有较好的识别率,基本实现跨数据集识别。     

深度学习在核物理中的应用

深度学习是目前最好的模式识别工具，预期会在核物理领域帮助科学家从大量复杂数据中寻找与某些物理最相关的特征。本文综述了深度学习技术的分类，不同数据结构对应的最优神经网络架构，黑盒模型的可解释性与预测结果的不确定性。介绍了深度学习在核物质状态方程、核结构、原子核质量、衰变与裂变方面的应用，并展示如何训练神经网络预测原子核质量。结果发现使用实验数据训练的神经网络模型对未参与训练的实验数据拥有良好的预测能力。基于已有的实验数据外推，神经网络对丰中子的轻原子核质量预测结果与宏观微观液滴模型有较大偏离。此区域可能存在未被宏观微观液滴模型包含的新物理，需要进一步的实验数据验证。     

应用神经网络的复杂物体三维测量

将神经网络引入基于结构光投影的复杂物体三维面形测量。在测量过程中,利用神经网络强大的函数逼近能力,得到离散条纹图的连续逼近函数,从中解出物体的相位分布信息,获得物体的三维面形分布。应用神经网络方法,在结构光投影条件下,只需要获取一幅条纹图,便可以完成复杂物体的三维面形测量。该方法相比传统的傅里叶变换轮廓术,不存在滤波操作,不会在测量过程中丢失被测物体的高频分量,具有更高的空间带宽积和灵敏度,能准确测量出复杂物体的细节,更加适用于恢复复杂物体的三维面形。并且该方法在条纹图存在阴影的情况下与傅里叶变换轮廓术相比,能更好地提取出物体的相位信息,恢复物体的三维面形。模拟及实验均验证了该方法的可行性。     

基于神经网络的托卡马克能量约束时间预测

使用机器学习理论中的神经网络方法,根据通用逼近原理对能量约束时间的复杂函数进行逼近,采用托卡马克装置的典型实验数据,设计一种组合结构的神经网络。通过大量的调参试验,给出一套性能最好的参数组合,并与传统幂指数形式的多元线性回归方法进行准确性和数据集迁移能力的比较。结果表明：神经网络模型对于能量约束时间的预测准确率更高,回归性能更好,且具有一定的抗噪声能力,更适合作为能量约束时间的定标或预测工具。     

Time series-based bifurcation diagram reconstruction

We consider the problem of reconstructing bifurcation diagrams (BDs) of maps using time series. This study goes along the same line of ideas presented by Tokunaga et al. [Physica D 79 (1994) 348] and Tokuda et al. [Physica D 95 (1996) 380]. The aim is to reconstruct the BD of a dynamical system without the knowledge of its functional form and its dependence on the parameters. Instead, time series at different parameter values, assumed to be available, are used. A three-layer fully-connected neural network is employed in the approximation of the map. The task of the network is to learn the dynamics of the system as function of the parameters from the available time series. We determine a class of maps for which one can always find a linear subspace in the weight space of the network where the network’s bifurcation structure is qualitatively the same as the bifurcation structure of the map. We discuss a scheme in locating this subspace using the time series. We further discuss how to recognize time series generated by this class of maps. Finally, we propose an algorithm in reconstructing the BDs of this class of maps using predictor functions obtained by neural network. This algorithm is flexible so that other classes of predictors, apart from neural networks, can be used in the reconstruction.     

Sensor and Actuator Fault Diagnosis for Robot Joint Based on Deep CNN

This paper proposes a data-driven method-based fault diagnosis method using the deep convolutional neural network (DCNN). The DCNN is used to deal with sensor and actuator faults of robot joints, such as gain error, offset error, and malfunction for both sensors and actuators, and different fault types are diagnosed using the trained neural network. In order to achieve the above goal, the fused data of sensors and actuators are used, where both types of fault are described in one formulation. Then, the deep convolutional neural network is applied to learn characteristic features from the merged data to try to find discriminative information for each kind of fault. After that, the fully connected layer does prediction work based on learned features. In order to verify the effectiveness of the proposed deep convolutional neural network model, different fault diagnosis methods including support vector machine (SVM), artificial neural network (ANN), conventional neural network (CNN) using the LeNet-5 method, and long-term memory network (LTMN) are investigated and compared with DCNN method. The results show that the DCNN fault diagnosis method can realize high fault recognition accuracy while needing less model training time.