A Microassembled Low-Profile Three-Dimensional Microelectrode Array for Neural Prosthesis Applications

This paper describes the design and micro- assembly process of a low-profile 3-D microelectrode array for mapping the functional organization of targeted areas of the central nervous system and for possible application in neural prostheses. The array consists of multiple planar complimentary metal-oxide-semiconductor stimulating probes and 3-D assembly components. Parylene-encapsulated gold beams supported by etch-stopped silicon braces allow the backends of the probes to be folded over to reduce the height of the array above the cortical surface. A process permitting parylene to be used at wafer level with bulk-silicon wet release has been reported. Spacers are used to fix the microassembled probes in position and are equipped with interlocking structures to facilitate the assembly process and increase yield. Four-probe 256-site 3-D arrays operate from plusmn5 V with an average per-channel power dissipation of 97 muW at full range stimulation with pulse widths of 100 mus at 500-Hz frequency. Thirty-two sites can be stimulated simultaneously with maximum currents of plusmn127 muA and a current resolution of plusmn1 muA. The microassembly techniques allow a variety of 3-D microstructures to be created from planar components.     

A Novel Self-Organizing Neural Network for Motion Segmentation

Many computer vision techniques, above all for structure from motion problems, require a segmentation of the images captured by one or more cameras. This paper deals with the segmentation based on the motion information, but can be easily extended to other cases (color, texture and so on). A new neural network, the EXIN Segmentation Neural Network (EXIN SNN) is here introduced. It is incremental, self-organizing and considers its task as the solution of a pattern recognition problem. This original approach overcomes the limits of the traditional segmentation techniques, namely the need of a spatial support for the image objects and the translation parallel to the image plane for the objects in the scene. Examples are given both for synthetic and real images.     

A Novel Two-stage Learning Pipeline for Deep Neural Networks

In this work, a training method was proposed for Deep Neural Networks (DNNs) based on a two-stage structure. Local DNN models are trained in all local machines and uploaded to the center with partial training data. These local models are integrated as a new DNN model (combination DNN). With another DNN model (optimization DNN) connected, the combination DNN forms a global DNN model in the center. This results in greater accuracy than local DNN models with smaller amounts of data uploaded. In this case, the bandwidth of the uploaded data is saved, and the accuracy is maintained as well. Experiments are conducted on MNIST dataset, CIFAR-10 dataset and LFW dataset. The results show that with less training data uploaded, the global model produces greater accuracy than local models. Specifically, this method focuses on condition of big data.     

A Novel Continuous Forward Algorithm for RBF Neural Modelling

A continuous forward algorithm (CFA) is proposed for nonlinear modelling and identification using radial basis function (RBF) neural networks. The problem considered here is simultaneous network construction and parameter optimization, well-known to be a mixed integer hard one. The proposed algorithm performs these two tasks within an integrated analytic framework, and offers two important advantages. First, the model performance can be significantly improved through continuous parameter optimization. Secondly, the neural representation can be built without generating and storing all candidate regressors, leading to significantly reduced memory usage and computational complexity. Computational complexity analysis and simulation results confirm the effectiveness     

一种神经网络辨识的混合学习算法

文章提出了一种神经网络辨识的混合学习算法。采用具有递阶结构的遗传算法来获得神经网络拓扑结构和连接权值的全局次优解,之后由BP算法来进一步调整神经网络的连接权值,从而实现神经网络的自动优化设计。仿真结果表明,所得的神经网络结构简单、精度高,并具有良好的泛化能力。     

A Novel Self-Creating Neural Network for Learning Vector Quantization

This paper presents a novel self-creating neural network scheme which employs two resource counters to record network learning activity. The proposed scheme not only achieves the biologically plausible learning property, but it also harmonizes equi-error and equi-probable criteria. The training process is smooth and incremental: it not only avoids the stability-and-plasticity dilemma, but also overcomes the dead-node problem and the deficiency of local minimum. Comparison studies on learning vector quantization involving stationary and non-stationary, structured and non-structured inputs demonstrate that the proposed scheme outperforms other competitive networks in terms of quantization error, learning speed, and codeword search efficiency.     

全局优化RBF网络的一种新算法

神经网络的输入变量、隐含层结点以及中心的选择对模型的性能都有重大的影响,以前的研究一般只考虑优化网络的参数或其结点数。为解决这个问题,提出了一种新的全局优化算法来自动选择RBF神经网络的输入变量和结点数目,并同时优化其参数。在提出的算法中,RBF网络的结点数目、输入变量的选择和参数都采用二进制编码,并用遗传算法来优化。为提高算法的性能和收敛速度,在遗传算法优化的同时引入了一种高性能的基于梯度的局部搜索算子(结构化的非线性参数优化方法)来优化RBF网络中的参数。Box-Jenkins煤气炉标准时间序列的预测问题被用来检验算法的性能。实验结果表明,提出的算法可以得到非常"紧凑"的RBF网络,且其性能优于其他一些算法。     

一种新的双权值前向神经网络学习算法

提出了一种新的基于遗传算法和误差反向传播的双权值神经网络学习算法,同时确定核心权值、方向权值以及幂参数、学习率等参数,通过适当地调节这些参数,可以实现尽可能多种不同超曲面的特性以及起到加快收敛的效果。并通过对实际的模式分类问题的仿真,将文中的方法与带动量项BP算法、CSFN等算法进行了比较,验证了其有效性。实验结果表明所提出的方法具有分类准确率高、收敛速度快的优点。     

一种适用于电力系统谐波分析的神经网络模型

提出一种适用于电力系统谐波分析的神经网络模型,为求解该神经网络权值参数.推导了基于非线性优化思想的权值参数自适应学习算法.该算法同步调整隐层神经元和输出层神经元权值参数,以优化神经网络能量函数.为了验证算法的有效性,应用该算法对电力系统谐波进行仿真实倒研究.仿真结果表明,利用该算法能以非常高的精度同步求解电力系统基波频率以及各次谐波的幅值和相位.     

A Novel Weightless Artificial Neural Based Multi-Classifier for Complex Classifications

Artificial neural systems in general and weightless systems in particular, have traditionally struggled in performance terms when confronted with problem domains such as possessing a large number of independent pattern classes and pattern classes with non-standard distributions. A multi-classifier is proposed which explores problem domains with a large number of independent pattern classes typically found in forensic and security databases. Specifically, the multi-classifier system is demonstrated on the exemplar of fingerprint identification problem typical to forensic, biometric, and security. Furthermore, the multi-classifier is able to provide a reasonable solution to benchmark problems from medicinal and physical (science) fields, which are determining the health, state of thyroid glands and determining whether or not there is a structure in the ionosphere, respectively.     

一种新的神经网络结构及其学习算法

提出了一个基于模糊集理论的新的神经网络结构及其学习算法。这种神经网络是对标准的BP网络修改后构成的(记作FIBP)。并通过几个实例仿真验证其有效性。实验结果表明,当用FIBP网络解决动态的、高度非线性的函数逼近时,其学习速度比BP网络快,精度高而且泛化能力高。     

A Novel Approach Solving for Linear Matrix Inequalities Using Neural Networks

Linear matrix inequalities (LMIs) play avery important role in postmodern control by providinga framework that unifies many concepts. While numerouspapers have appeared cataloging applications of LMIsto control system analysis and design, there have beenfew publications in the literature describing thenumerical solution of these problems. Specially, neural network processing has rarely been used to solve those problems.This paper attempts topropose a new approach to solving a class of LMIsusing recurrent neural networks. The nature ofparallel and distributed processing renders thesenetworks, which possess the computational advantages overthe traditional sequential algorithms in real-timeapplications. The proposed networks are proven to be largelyasymptotical and capable of solving LMIs.Some illustrative examples are provided todemonstrate the proposed results.     

A Novel Attention-based Global and Local Information Fusion Neural Network for Group Recommendation

Due to the popularity of group activities in social media, group recommendation becomes increasingly significant. It aims to pursue a list of preferred items for a target group. Most deep learning-based methods on group recommendation have focused on learning group representations from single interaction between groups and users. However, these methods may suffer from data sparsity problem. Except for the interaction between groups and users, there also exist other interactions that may enrich group representation, such as the interaction between groups and items. Such interactions, which take place in the range of a group, form a local view of a certain group. In addition to local information, groups with common interests may also show similar tastes on items. Therefore, group representation can be conducted according to the similarity among groups, which forms a global view of a certain group. In this paper, we propose a novel global and local information fusion neural network (GLIF) model for group recommendation. In GLIF, an attentive neural network (ANN) activates rich interactions among groups, users and items with respect to forming a group′s local representation. Moreover, our model also leverages ANN to obtain a group′s global representation based on the similarity among different groups. Then, it fuses global and local representations based on attention mechanism to form a group′s comprehensive representation. Finally, group recommendation is conducted under neural collaborative filtering (NCF) framework. Extensive experiments on three public datasets demonstrate its superiority over the state-of-the-art methods for group recommendation.

     

A Gas-Actuated Anthropomorphic Prosthesis for Transhumeral Amputees

This paper presents the design of a gas-actuated anthropomorphic arm prosthesis with 21 degrees of freedom and nine independent actuators. The prosthesis utilizes the monopropellant hydrogen peroxide as a gas generator to power nine pneumatic type actuators. Of the nine independent actuators, one provides direct- drive actuation of the elbow, three provide direct-drive actuation for the wrist, and the remaining five actuate an underactuated 17 degree of freedom hand. This paper describes the design of the prosthesis, including the design of small-scale high-performance servovalves, which enable the implementation of the monopropellant concept in a transhumeral prosthesis. Experimental results are given characterizing both the servovalve performance and the force and/or motion control of various joints under closed-loop control.     

A Novel Method to Handle Inequality Constraints for Convex Programming Neural Network

By modifying the multipliers associated with inequality constraints, we can directly solve convex programming problem without nonnegative constraints of the multipliers associated with inequality constraints, hence it is no longer necessary to convert the inequality constraints into the equality constraints by using the slack variables. With this technique, the neural network to solve convex programming problem is constructed, and its stability is analyzed rigorously. Simulation shows that this method is feasible.     

A Novel Pruning Algorithm for Optimizing Feedforward Neural Network of Classification Problems

Optimizing the structure of neural networks is an essential step for the discovery of knowledge from data. This paper deals with a new approach which determines the insignificant input and hidden neurons to detect the optimum structure of a feedforward neural network. The proposed pruning algorithm, called as neural network pruning by significance (N2PS), is based on a new significant measure which is calculated by the Sigmoidal activation value of the node and all the weights of its outgoing connections. It considers all the nodes with significance value below the threshold as insignificant and eliminates them. The advantages of this approach are illustrated by implementing it on six different real datasets namely iris, breast-cancer, hepatitis, diabetes, ionosphere and wave. The results show that the proposed algorithm is quite efficient in pruning the significant number of neurons on the neural network models without sacrificing the networks performance.     

一种新兴的数据挖掘方法:神经规则法

该文详细阐述了一种新兴的数据挖掘方法——神经规则法。文中先后介绍了神经网络的训练、神经网络的剪枝以及神经规则的提取。最后用一个实例具体解释了使用该方法来挖掘分类规则的过程。     

高速视频记录应用的一种解决方案

在PC平台上进行高速数据记录,由于受局部总线速度和文件系统的影响,速度很难超过每秒20M字节。我们设计了一种新的方案,总线采用时钟频率为66MHz,数据位宽为64bits的PCI总线,提高了总线速度。采用Adaptec公司的ASPI指令直接控制SCSI接口卡读写硬盘,摆脱了文件系统对写入速度的影响。记录媒质是由两块SCSI硬盘组成的RAID0级阵列,提高存储媒体的物理写入速度。实验结果证明,存储速度得到很大提高,稳定速度达到每秒100M字节以上。     

一种改进的神经网络非线性预测控制

从建立神经网络非线性预测模型出发,针对BP网络存在收敛速度慢,容易陷入局部最小的缺点,该文在BFGS拟牛顿法的基础上,提出了一种基于并行拟牛顿优化算法的并行拟牛顿神经网络。该并行拟牛顿优化算法采用两个含有不同参数的拟牛顿校正公式,在每次迭代过程中,利用这两个不同的校正公式得到相应的搜索方向,并通过不精确搜索法求取最优步长,最后根据一性能指标取最优的一个搜索方向和相应的步长对网络各层之间的权值进行修正。Matlab仿真结果表明,同BP神经网络和BFGS拟牛顿神经网络相比,该神经网络具有收敛速度快、模型精度高的特点,更适合于实时非线性控制。     

Novel Neuronal Activation Functions for Feedforward Neural Networks

Feedforward neural network structures have extensively been considered in the literature. In a significant volume of research and development studies hyperbolic tangent type of a neuronal nonlinearity has been utilized. This paper dwells on the widely used neuronal activation functions as well as two new ones composed of sines and cosines, and a sinc function characterizing the firing of a neuron. The viewpoint here is to consider the hidden layer(s) as transforming blocks composed of nonlinear basis functions, which may assume different forms. This paper considers 8 different activation functions which are differentiable and utilizes Levenberg-Marquardt algorithm for parameter tuning purposes. The studies carried out have a guiding quality based on empirical results on several training data sets.