Improved neural network model for reverse engineering

While conventional engineering transforms engineering concepts into real parts, in reverse engineering real parts are transformed into engineering models. The construction of a surface from three-dimensional (3D) measuring data points is an important problem in reverse engineering. This paper presents a reconstruction method for the sculptured surfaces from the 3D measuring data points. The surface reconstruction scheme is presented based on a neural network. The reconstruction of the existing surfaces is realized by training the network. A series of measuring points from existing sculptured surfaces is used as a training set. Once the neural network has been trained, it serves as a geometric model to generate all the points that are needed. However, the learning rate for the neural network is relatively slow, and the learning accuracy is often unacceptably low. In this paper, to improve the performance of the neural network, a pre-processor is proposed before the input layer. The pre-processor maps the input into the larger space by generating a set of linearly independent values. The effect of the pre-processor is to increase modelling accuracy, and reduce learning time. Based on this method, experimental results are given to show that the reconstructed surfaces are faithful to the original data points. The proposed scheme is useful for regular or irregular digitized data.     

Experimental exploration on loss surface of deep neural network

The loss function of the deep neural network is high dimensional, nonconvex and complex. So far, the geometric properties of the loss surface of the neural network have not been well understood. Different from most theoretical studies on the loss surface, this article makes the experimental exploration on the loss surface of the deep neural network, including trajectories of various adaptive optimization algorithms, the Hessian matrix of the loss function of the deep neural network, the curvature of the loss surface along the trajectories of the various adaptive optimization algorithms. It is found that the gradient direction of the adaptive optimization algorithms is almost perpendicular to the direction of the maximum curvature of the loss surface, while the gradient directions of the stochastic gradient descent (SGD) algorithm do not have such a rule. The Hessian matrix of the loss surface along the trajectory of the optimization algorithm is degraded, which is inconsistent with the hypothetical that nonsingular of the Hessian matrix in many theoretical studies of deep learning. Besides, this article proposes a new ensemble learning method of the neural network based on the scaling invariance of the ReLu neural network and mode connectivity.     

基于模糊学习算法的空调机组性能模拟

基于模糊熵准则和误差平方和准则建立了模糊学习算法,基于该模糊学习算法,应用BP神经网络对柜式空调机组的性能进行了模拟.结果表明,与传统的基于误差平方和准则的学习算法相比,采用模糊学习算法可以大大简化网络结构,有效提高模拟精度和效率.     

Artificial neural networks for predicting the macromechanical behaviour of ceramic-matrix composites

In ceramic-matrix composites (CMCs) a weak fibre/matrix interface is required to achieve satisfactory toughening so that the composite exhibits damage tolerant characteristics. Due to the presence of such a weak interface, debonding and sliding occur at the interface making the mechanics of the material very complex. As a result, developing analytical models for simulating the macromechanical behaviour of these composites is extremely difficult and necessitates simplifying assumptions compromising accuracy. In the present paper, a novel approach to modelling the macromechanical behaviour of CMCs, using the artificial neural network (ANN) approach has been presented. The ability of neural networks in learning the complex multi-parametric interaction among the various microstructural parameters has been demonstrated with an example of SiC/SiC ceramic composite. An artificial neural network has been used to postulate the macromechanical behaviour of SiC (matrix)/SiC (fibre) composite. The training examples for the network have been generated through an accurate micromechanical finite element analysis that models the interfacial debonding and sliding realistically. The network learning is demonstrated and the network is validated by asking it to predict the behaviour of the composite for new specimens. Various stages in the development of ANN such as the preparation of training set, selection of a network configuration, training of the net and a testing scheme, etc, have been addressed at length in this paper.     

Expert Diagnosing system for a Rotation Mechanism Based on a Neural Network

By combining the artificial neural network with the rule reasoning expert system,an expert diagnosing system for a rotation mechanism was established.This expert system takes advantage of both a neural network and a rule reasoning expert system;it can also make use of all kinds of knowledge in the repository to diagnose the fault with the positive and negative mixing reasoning mode.The binary system was adopted to denote all kinds of fault in a rotation mechanism.The neural networks were trained with a random parallel algorithm (Alopex).The expert system overcomes the self-learning difficulty of the rule reasoning expert system and the shortcoming of poor system control of the neural network.The expert system developed in this paper has power ful diagnosing ability.     

卷积神经网络在印刷品缺陷检测的应用

目的 为了改善传统机器检测印刷产品缺陷存在误费率高的不足。方法 提出以卷积神经网络为控制核心的印刷品缺陷检测系统。设计可在实际检测中应用的卷积神经网络,设计在线印刷质量检测系统的硬件结构。结果 对结构相同而训练次数、学习率不同的卷积神经网络进行了缺陷检测的性能对比,验证了该卷积神经网络在学习率小于0.01时,可以获得较好的识别效果;在学习率大于0.05时,网络不容易收敛。网络训练次数越多,精度越高,相应的训练时间也较长。在满足快速性和精确度的条件下,确定了适应某印刷品的缺陷检验网络训练次数为50,学习率为0.005,此时的识别率为90%。结论 经过实验证明,该检测系统具有良好的缺陷识别能力,缺陷类型的分类准确率较高。该系统具有一定的实用价值。     

Rule Extraction from Trained Artifical Neural Networks

The neural network rule extraction problem is aimed at obtaining rules from an arbitrarily trained artificial neural network. Recently there have been several approaches to rule extraction. Approaches to rule extraction implement a priori knowledge of data or rule requirements into neural networks before the rules are extracted. Although this may lead to a simplified final phase of acquitting the rules from particular type of neural networks, it limits the methodologies for general-purpose use. This article approaches the neural network rule extraction problem in its essential and general form. Preference is given to multilayer perceptron networks (MLP networks) due to their universal approximation capabilities. The article establishes general theoretical grounds for rule extraction from trained artificial neural networks and further focuses on the problem of crisp rule extraction. The problem of crisp rule extraction from trained MLP networks is first approached on theoretical level. Present ed theoretical results state conditions guaranteeing equivalence between classification by an MLP network and crisp logical formalism. Based on the theoretical results an algorithm for crisp rule extraction, independent of training strategy, is proposed. The rule extraction algorithm can be used even in cases where the theoretical conditions are not strictly satisfied; by offering an approximate classification. An introduced rule extraction algorithm is experimentally demonstrated.

     

Incorporating pattern recognition capability in a flexible vibratory bowl feeder using a neural network

Programmable parts feeders that can orientate most of the parts of one or more part families, with short changeover times from one part to the next, are highly sought after in batch production. This study investigates a suitable neural-network-based pattern recognition algorithm for the recognition of parts in a programmable vibratory bowl feeder. Three fibre-optic sensors were mounted on a vibratory bowl feeder to scan the surface of each feeding part. The scanned signatures were used as the input for the different neural network models. The performances of ARTMAP, ART2 and backpropagation neural network models were compared. The results showed that, among the three models, ARTMAP is deemed to be superior, based on the criteria of learning speed, high generalization and flexibility. The better performance obtained with the ARTMAP neural network is mainly the result of its online training and supervised learning capabilities.     

A two‐stage neural network prediction of chronic kidney disease

Accurate detection of chronic kidney disease (CKD) plays a pivotal role in early diagnosis and treatment. Measured glomerular filtration rate (mGFR) is considered the benchmark indicator in measuring the kidney function. However, due to the high resource cost of measuring mGFR, it is usually approximated by the estimated glomerular filtration rate, underscoring an urgent need for more precise and stable approaches. With the introduction of novel machine learning methodologies, prediction performance is shown to be significantly improved across all available data, but the performance is still limited because of the lack of models in dealing with ultra‐high dimensional datasets. This study aims to provide a two‐stage neural network approach for prediction of GFR and to suggest some other useful biomarkers obtained from the blood metabolites in measuring GFR. It is a composite of feature shrinkage and neural network when the number of features is much larger than the number of training samples. The results show that the proposed method outperforms the existing ones, such as convolutionneural network and direct deep neural network.     

Improved probabilistic neural network algorithm for chemical sensor array pattern recognition

An improved probabilistic neural network (IPNN) algorithm for use in chemical sensor array pattern recognition applications is described. The IPNN is based on a modified probabilistic neural network (PNN) with three innovations designed to reduce the computational and memory requirements, to speed training, and to decrease the false alarm rate. The utility of this new approach is illustrated with the use of four data sets extracted from simulated and laboratory-collected surface acoustic wave sensor array data. A competitive learning strategy, based on a learning vector quantization neural network, is shown to reduce the storage and computation requirements. The IPNN hidden layer requires only a fraction of the storage space of a conventional PNN. A simple distance-based calculation is reported to approximate the optimal kernel width of a PNN. This calculation is found to decrease the training time and requires no user input. A general procedure for selecting the optimal rejection threshold for a PNN-based algorithm using Monte Carlo simulations is also demonstrated. This outlier rejection strategy is implemented for an IPNN classifier and found to reject ambiguous patterns, thereby decreasing the potential for false alarms.     

Biomedical Event Extraction Using a New Error Detection Learning Approach Based on Neural Network

Supervised machine learning approaches are effective in text mining, but their success relies heavily on manually annotated corpora. However, there are limited numbers of annotated biomedical event corpora, and the available datasets contain insufficient examples for training classifiers; the common cure is to seek large amounts of training samples from unlabeled data, but such data sets often contain many mislabeled samples, which will degrade the performance of classifiers. Therefore, this study proposes a novel error data detection approach suitable for reducing noise in unlabeled biomedical event data. First, we construct the mislabeled dataset through error data analysis with the development dataset. The sample pairs’ vector representations are then obtained by the means of sequence patterns and the joint model of convolutional neural network and long short-term memory recurrent neural network. Following this, the sample identification strategy is proposed, using error detection based on pair representation for unlabeled data. With the latter, the selected samples are added to enrich the training dataset and improve the classification performance. In the BioNLP Shared Task GENIA, the experiments results indicate that the proposed approach is competent in extract the biomedical event from biomedical literature. Our approach can effectively filter some noisy examples and build a satisfactory prediction model.     

改进的BP神经网络在风机故障诊断中的应用

采用动量法和学习速率自适应的改进BP神经网络建立风机故障诊断系统。在网络训练过程中分别采用标准训练样本和含有白噪声的训练样本来训练网络,使网络具有一定的容错性。最后通过仿真实验和风机的故障诊断实例表明：改进的BP神经网络减少训练次数,提高了学习效率,而且有效地抑制网络陷于局部极小,是风机故障诊断的有效方法。     

A Noise-Tolerant Hybrid Model of A Global and A Local Learning Module

Proposed is GLLL2, a hybrid architecture of a global and a local learning module, which learns default and exceptional knowledge respectively from noisy examples. The global learning module, which is a feedforward neural network, captures global trends gradually, while the local learning module stores local exceptions quickly. The latter module distinguishes noise from exceptions, and learns only exceptions, which makes GLLL2 noise-tolerant. The results of experiments show the process in which training examples are formed into default and exceptional knowledge, and demonstrate that the predictive accuracy, the space efficiency, and the training efficiency of GLLL2 is higher than those of each individual module.

     

Detecting process mean shift in the presence of autocorrelation: a neural-network based monitoring scheme

The main purpose of this paper is twofold: (1) to present a neural-network based methodology for monitoring process shift in the presence of autocorrelation; and (2) to demonstrate the power, the effectiveness, and the adaptability of this approach. The proposed neural network uses the effective and efficient extended delta-bar-delta learning rule and can be trained with the powerful back-propagation algorithm. The comparative study on AR(1) processes shows that the performance of this neural-network based monitoring scheme is superior to that of SCC, X, EWMA, EWMAST and ARMAST control charts in most instances. Moreover, the network output can also provide information about the shift magnitude. The study of run length distributions suggests that further improvement on designing such neural networks is possible. The adaptability of the neural-network approach is demonstrated through the flexible design of the training data set. To further improve run length properties under various shift magnitudes, alternative control heuristics are proposed.     

用BP神经网络诊断结构破损

提出了一种基于BP神经网络的结构破损诊断方法,该方法以结构残余力向量作为破损诊断的网络输入。对网络训练样本采用广义空间格点法进行了变换,从而较好地解决了由于系统响应样本在数据空间分布不均对网络收敛速度及网络诊断精度的影响问题。应用实例表明,本文方法能准确诊断结构破损位置与严重程度,是一种有效的结构破损诊断方法。     

Optical implementation of a translation-invariant second-order neural network for multiple-pattern classification

A novel approach to the optical implementation of second-order neural networks that can recognize multiple patterns is reported. The systems issues, especially the accuracy required for the weighted interconnections, are discussed for numeric character (0-9) recognition. It is shown that the accuracy of the weighted interconnections has a far greater influence on the network performance during training than on classification. To lessen the problem, we introduce an adaptive learning rule, whereby the optical power is adjusted during training. Finally, numeric character recognition using an experimental system with a liquid-crystal display is demonstrated.     

扩展双向联想记忆神经网络及其应用

提出了一种采用δ规则作为学习算法的扩展双向联想记忆神经网络模型，并从理论上证明了其稳定性。该模型克服了现有采用Ｈｅｂｂ规则作为学习算法的联想记忆神经网络对记忆模式有正交性要求和所模式吸引域小的不足。实验结果表明，其联想记忆能力优于目前现有的联想记忆网络。     

基于GA-BP神经网络的结构损伤位置识别

针对传统BP神经网络训练中存在的一些问题，提出了一种基于遗传算法(GA)-BP神经网络混合技术识别结构损伤位置的方法。该方法利用基因实数编码的遗传算法优化BP网络的结构及初始参数，从而大大提高了神经网络的训练精度。运用GA-BP网络与传统BP网络技术分别对两个算例进行了结构损伤定位的识别仿真，结果表明遗传BP稳定性好，精度高，对噪声有很好的鲁棒性，便于工程应用。     

多维线性模型参数估计的神经网络方法

本文提出一种用于多维线性模型（ＡＲ，ＡＲＭＡ）参数估计的神经网络方法和相应的递归预测误差算法。本文在分析多输入、单输出，含一个隐含和多层神经网络的输入输出关系的基础上，提出了首先将理想输出Ｘｉ进行预畸变（Ｆ（Ｘｔ））作为神经网络的训练目标。当神经网络训练完毕后，网络的连接权就是待估计的线性模型参数。本文提出方法的优点是网络结构简单，估计结果准确。仿真模拟结果表明，本文所提出的神经网络方法估计多维线性模型参数是有效的。     

An efficient gait recognition based on a selective neural network ensemble

The neural network ensemble is a learning paradigm where a collection of neural networks is trained for the same task. Generally, the ensemble shows better generalization performance than a single neural network. In this article, a selective neural network ensemble is applied to gait recognition. The proposed method selects some neural network based on the minimization of generalization error. Since the selection rule is directly incorporated into the cost function, we can obtain adequate component networks to constitute an ensemble. Experiments are performed with the NLPR database to show the performance of the proposed algorithm. © 2008 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 18, 237–241, 2008; Published online in Wiley InterScience (www.interscience.wiley.com).