Symbols Language Neural Networks Performance

An implementation of non-regular symbol manipulation with neural networks is presented. In particular, it is shown how a context-free language can be produced with neural networks. The rules of the language are stored as patterns in an attractor neural network. Another such network is used as a working memory, which can be enlarged without changing the production system itself. As a result, the competence of symbol manipulation with neural networks equals that of classical non-regular production systems. In actual behaviour (performance), however, there are differences between the systems, which shows the importance of implementation in the generation of rule-like behaviour.     

Combining Boolean Neural Architectures for Image Recognition

This paper presents a completely integrated Boolean neural architecture, where a selforganizing Boolean neural network (SOFT) is used as a front-end processor to a feedforward Boolean network based on goal-seeking principles (GSN f). This paper will evaluate the advantages of the integrated SOFT-GSN f over GSN f by showing its increased effectiveness in an optical character recognition task.     

Modular Neural Networks for Medical Prognosis: Quantifying the Benefits of Combining Neural Networks for Survival Prediction

This paper describes a medical application of modular neural networks (NNs) for temporal pattern recognition. In order to increase the reliability of prognostic indices for patients living with the acquired immunodeficiency syndrome (AIDS), survival prediction was performed in a system composed of modular NNS that classified cases according to death in a certain year of follow-up. The output of each NN module corresponded to the probability of survival in a given year. Inputs were the values of demographic, clinical and laboratory variables. The results of the modules were combined to produce survival curves for individuals. The NNs were trained by backpropagation and the results were evaluated in test sets of previously unseen cases. We showed that, for certain combinations of NN modules, the performance of the prognostic index, measured by the area under the receiver operating characteristic curve, was significantly improved (p 0.05). We also used calibration measurements to quantify the benefits of combining NN modules, and show why, when and how NNs should be combined for building prognostic models.     

基于WTD-AR谱和MEA-BPNN的轴承故障诊断方法

针对滚动轴承故障诊断模型在噪声干扰下鲁棒性能差的问题,提出一种基于小波阈值去噪（WTD）、AR谱和思维进化算法（MEA）优化反向传播神经网络（BPNN）的轴承故障诊断方法。以原始振动信号为输入,采用小波方法分解重构原始信号滤除高频噪声,然后采用Burg算法估计AR模型参数提取降噪信号功率谱特征,最后将特征向量与对应标签分别作为MEA-BPNN神经网络的输入、输出进行训练,最终实现诊断。将该方法与一些先进的人工神经网络诊断方法作比较,测试该诊断模型的性能。研究结果表明：WTD-AR谱-MEA-BPNN诊断模型能够有效降低轴承振动信号的噪声干扰,实现特征增强,分辨率更高;相较于传统神经网络训练速度更快,在更短时间内甄别故障类型且识别率高。     

On Combining Artificial Neural Nets

This paper reviews research on combining artificial neural nets, and provides an overview of, and an introduction to, the papers contained in this special issue, and its companion (Connection Science, 9, 1). Two main approaches, ensemble-based, and modular, are identified and considered. An ensemble, or committee, is made up of a set of nets, each of which is a general function approximator. The members of the ensemble are combined in order to obtain better generalization performance than would be achieved by any of the individual nets. The main issues considered here under the heading of ensemble-based approaches are a how to combine the outputs of the ensemble members, b how to create candidate ensemble members and c which methods lead to the most effective ensembles? Under the heading of modular approaches, we begin by considering a divide-and-conquer approach by which a function is automatically decomposed into a number of subfunctions which are treated by specialist modules. Other modular approaches are also identified and considered, for while the divide-and-conquer approach is designed to improve performance, the term modularity can be given a wider interpretation. The broadly defined topic of modularity includes the explicit decomposition of a task based on the designer's understanding, and the exploitation of specialist modules in order to accomplish tasks which could not be performed by a monolithic net.     

Modularity,Combining and Artificial Neural Nets

In this paper, the modular combination of artificial neural nets is considered. A modular approach to combining can be contrasted with an ensemble-based approach in that it implies individual modules, each responsible for some specialist aspect of a task, as opposed to each approximating the same function. It is possible to characterize modular systems in terms of (i) reasons for the task decomposition, (ii) the method for accomplishing the decomposition and (iii) the relationship between the modules. These characteristics are considered in brief outlines of the papers in the issue. Reasons for task decomposition include the exploitation of specialist capabilities of individual nets, performance improvement, and making the system easier to understand and modify. Task decomposition may be either automatic (based on the blind application of a data partitioning algorithm) or explicit (based on prior knowledge of the task or the specialist capabilities of the modules), and the relationship between the modules may be successive, cooperative or supervisory.     

A Neural Gripper for Arbitrary Object Grasping

This paper presents a two-stage neural system to determine the contact points between a three-fingered gripper and an object of arbitrary shape. In the first stage, a CCD camera captures the image of the object and such an image is transformed into a two-dimensional outline through a nearest neighbour algorithm. In the second phase, two neural networks, functioning in cascade, select three contact points in the outline. A competitive Hopfield neural network defines an approximate polygon considering a reduced number of boundary points of the original outline. Then, a supervised neural network, either a multi-layer perceptron or a radial basis function (RBF) network, find the contact points. The experiments suggest that the RBF network trained by the global ridge regression method is suitable for on-line applications and presents the best overall performance in terms of accuracy and robustness to noise. Moreover, this method is able to find correctly the contact points for objects of arbitrary shapes.     

Hybridization and Specialization of Real-time Recurrent Learning-based Neural Networks

In this article, three different methods for hybridization and specialization of real-time recurrent learning (RTRL)-based neural networks (NNs) are presented. The first approach consists of combining recurrent networks with feedforward networks. The second approach continues with the combination of multiple recurrent NNs. The last approach introduces the combination of connectionist systems with instructionist artificial intelligence techniques. Two examples are added to demonstrate properties and advantages of these techniques. The first example is a process diagnosis task where a hybrid NN is connected to a knowledge-based system. The second example is a NN consisting of different recurrent modules that is used to handle missing sensor data in a process modelling task.     

Character Recognition Digit Recognition Pattern Recognition Spatiotemporal Neural Networks Modular Networks Segmentation Problem

We describe an alternate approach to visual recognition of handwritten words, wherein an image is converted into a spatio-temporal signal by scanning it in one or more directions, and processed by a suitable connectionist network. The scheme offers several attractive features including shift-invariance, explication of local spatial geometry along the scan direction, a significant reduction in the number of free parameters, the ability to process arbitrarily long images along the scan direction, and a natural framework for dealing with the segmentation/recognition dilemma. Other salient features of the work include the use of a modular and structured approach for network construction and the integration of connectionist components with a procedural component to exploit the complementary strengths of both techniques. The system consists of two connectionist components and a procedural controller. One network concurrently makes recognition and segmentation hypotheses, and another performs refined recognition of segmented characters. The interaction between the networks is governed by the procedural controller. The system is tested on three tasks: isolated digit recognition, recognition of overlapping pairs of digits and recognition of ZIP codes.     

Combining Neural Network Forecasts on Wavelet-transformed Time Series

We discuss a simple strategy aimed at improving neural network prediction accuracy, based on the combination of predictions at varying resolution levels of the domain under investigation (here: time series). First, a wavelet transform is used to decompose the time series into varying scales of temporal resolution. The latter provides a sensible decomposition of the data so that the underlying temporal structures of the original time series become more tractable. Then, a dynamical recurrent neural netork is trained on each resolution scale with the temporal-recurrent backpropagation algorithm. By virtue of its internal dynamic, this general class of dynamic connections network approximates the underlying law governing each resolution level by a system of non-linear difference equations. The individual wavelet scale forecasts are afterwards recombined to form the current estimate. The predictive ability of this strategy is assessed with the sunspot series.     

Connectionism;Neural Networks;Learning;Ambiguities;Reading

After a brief review of the different types and causes of ambiguous training data and the problems of learning from such data, a class of multi-target models are presented which suggest that neural networks are even better at solving these problems than previously realized. They are able to learn which non-ambiguous subset of a larger ambiguous set of training data best captures any underlying regularities in that data and hence optimize generalization while minimizing the problems of overtraining. It is also shown how the deliberate generation of ambiguous training data can begin to solve some of the longstanding representational problems of mapping time sequences, such as the alignment problem for reading and spelling. The general ideas are illustrated throughout with the well-known problem of tex-to-phoneme conversion, and detailed results of a range of neural network simulations are presented.     

Making a Low-dimensional Representation Suitable for Diverse Tasks

We introduce a new approach to the training of classifiers for performance on multiple tasks. The proposed hybrid training method leads to improved generalization via a better low-dimensional representation of the problem space. The quality of the representation is assessed by embedding it in a two-dimensional space using multi-dimensional scaling, allowing a direct visualization of the results. The performance of the approach is demonstrated on a highly non-linear image classification task.     

基于神经网络的基板图像识别软件设计与实现

在分析了现有技术的基础上,提出了一种基于神经网络的基板图像识别体系,并针对某型引线键合机,设计并实现了基于该体系的识别软件实验结果表明：该软件使用简单,具有较好的识别能力和学习能力。     

Evolution and Development of Modular Control Architectures for 1D Locomotion in Six-legged Animats

An evolutionary approach is used to design neural control architectures for virtual sixlegged animats. Using a geometry-oriented variation of the cellular encoding scheme and syntactic constraints that reduce the size of the genetic search space, the developmental programs of straight locomotion controllers are first evolved. One such controller is then included as the first module in a larger architecture, in which a second neural module is evolved and develops connections to the first one, so as to set locomotion on or offaccording to sustained or instantaneous external control signals. Such an incremental approach should prove useful to the automatic design of relatively complex control architectures that might, in particular, implement some cognitive abilities over and above mere stimulus-response mechanisms.     

基于BP神经网络和D-S证据理论的刀具磨损监测方法

将BP神经网络和D-S证据理论相结合的方法运用于刀具磨损监测中,采用小波包分解法对刀具磨损过程中产生的声发射信号进行特征提取,构建特征向量,利用BP神经网络识别判断刀具磨损状态;通过BP神经网络的输出结果和训练误差计算D-S证据理论的基本概率赋值,并用D-S证据理论对BP神经网络的识别结果进行决策级融合。实验结果表明:该方法避免了神经网络识别时的误诊,提高了整个刀具磨损监测系统识别的准确性和可靠性。     

强力旋压过程中缺陷诊断的神经网络方法

为了解决强旋过程中出现的缺陷问题，应用随机信号技术的神经网络技术，建立缺陷检测系统，检测结果与实际情况符合。     

Neural network control strategies for low specification servo actuators

The paper considers neural network based control techniques as a means of obtaining improved dynamic performance from low specification dc motor servo drives. Such motors are low cost but exhibit large torque variations, non-linearity and possess high hysterisis. The use of neural network methods is demonstrated to offer an improvement compared with a conventional approach using a classical PID technique. Low performance motors of the type considered when combined with a neural network controller achieved a level of performance suitable for automatic handling systems and automated guided vehicles (AGV).     

Speaker Normalization and Model Selection of Combined Neural Networks

This paper introduces bootstrap error estimation for automatic tuning of parameters in combined networks, applied as front-end preprocessors for a speech recognition system based on hidden Markov models. The method is evaluated on a large-vocabulary (10 000 words) continuous speech recognition task. Bootstrap estimates of minimum mean squared error allow selection of speaker normalization models improving recognition performance. The procedure allows a flexible strategy for dealing with inter-speaker variability without requiring an additional validation set. Recognition results are compared for linear, generalized radial basis functions and multi-layer perceptron network architectures.     

Retroactive interference in neural networks and in humans: The effect of pattern-based learning

Catastrophic interference is addressed as a problem that arises from pattern-based learning algorithms. As such, it is not limited to artificial neural networks but can be demonstrated in human subjects in so far as they use a pattern-based learning strategy. The experiment tests retroactive interference in humans learning lists of consonant-vowel-consonant nonsense syllable pairs. Results show significantly more interference for subjects learning patterned lists than subjects learning arbitrarily paired lists. To examine how different learning strategies depend on the structure of the learning task, a mixture-of-experts neural network model is presented. The results show how these strategies may interact to give rise to the results seen in the human data.