Minimally Connective,Auto-associative,Neural Networks

Classic barriers to using auto-associative neural networks to model mammalian memory include the unrealistically high synaptic connectivity of fully connected networks, and the relative paucity of information that has been stored in networks with realistic numbers of synapses per neuron and learning rules amenable to physiological implementation. We describe extremely large, auto-associative networks with low synaptic density. The networks have no direct connections between neurons of the same layer. Rather, the neurons of one layer are 'linked' by connections to neurons of some other layer. Patterns of projections of one layer on to another which form projective planes, or other cognate geometries, confer considerable computational power an the network.     

Self-organization of Multiple Winner-take-all Neural Networks

In this paper, analysis of the information content of discretely firing neurons in unsupervised neural networks is presented, where information is measured according to the network's ability to reconstruct its input from its output with minimum mean square Euclidean error. It is shown how this type of network can self-organize into multiple winner-take-all subnetworks, each of which tackles only a low-dimensional subspace of the input vector. This is a rudimentary example of a neural network that effectively subdivides a task into manageable subtasks.     

Connectionist learning: A comparison of neural networks and an optical thin-film multilayer model

Current work on connectionist models has been focused largely on artificial neural networks that are inspired by the networks of biological neurons in the human brain. However, there are also other connectionistarchitectures that differ significantly from this biological exemplar. We proposed a novel connectionist learning architecture inspired by the physics associated with optical coatings of multiple layers of thin-films in a previous paper (Li and Purvis 1999, Annals of Mathematics and Artificial Intelligence, 26: 1-4). The proposed model differs significantly from the widely used neuron-inspired models. With thin-film layer thicknesses serving as adjustable parameters (as compared with connection weights in a neural network) for the learning system, the optical thin-film multilayer model (OTFM) is capable of approximating virtually any kind of highly nonlinear mappings. The OTFM is not a physical implementation using optical devices. Instead, it is proposed as a new connectionist learning architecture with its distinct optical properties as compared with neural networks. In this paper we focus on a detailed comparison of neural networks and the OTFM (Li 2001, Proceedings ofINNS-IEEE International Joint Conference on Neural Networks, Washington, DC, pp. 1727-1732). We describe the architecture of the OTFM and show how it can be viewed as a connectionist learning model. We then present experimental results on solving a classification problem and a time series prediction problem that are typical of conventional connectionist architectures to demonstrate the OTFM's learning capability.     

Ti-17合金本构关系的人工神经网络模型

开发了一个基于神经网络的Ｔｉ１７ 合金的本构关系模型。首先利用ＴｈｅｒｍｅｃｍａｓｔｏｒＺ 型热模拟机等温压缩Ｔｉ１７ 合金, 研究在不同变形温度、变形程度和应变速率等工艺参数条件下流动应力的变化情况。然后用实验所得的热变形工艺参数与性能间的数据训练人工神经网络。训练结束后的神经网络变成为一个知识基的本构关系模型。利用该模型预测的流动应力的值与实验结果间的误差较小。     

Modeling the compressive strength of molasses–cement sand system using design of experiments and back propagation neural network

Molasses, an eco-friendly and relatively cheap binder may be used as a substitute for chemical binders. For commercial exploitation of the molasses–cement sand system it is essential to generate models for predicting the properties of the sand mix from the composition. Central composite design is used to develop regression equations for predicting compressive strength of the sand mix when molasses is varied between 5.5% and 7.5% and cement between 2% and 4%. Though central composite design is an effective tool for studying the complex effects of number of independent variables on response factor it has quite a few limitations. Back propagation neural network is not only capable of modeling highly non-linear relationship using dispersed data in the solution domain but has a few advantages over the central composite design. But one of the major drawbacks of this network is that no theoretical basis exists to determine the number of hidden layers and number of neurons therein. Different configurations of BPNN have great effects on the predicted results. Back propagation neural networks of different configurations are trained. Results obtained form these networks are analyzed and compared with those obtained form regression equations and experiments. Guidelines for selecting the effective configuration of back propagation networks are proposed.     

Modeling development of natural multi-sensory integration using neural self-organisation and probabilistic population codes

Humans and other animals have been shown to perform near-optimally in multi-sensory integration tasks. Probabilistic population codes (PPCs) have been proposed as a mechanism by which optimal integration can be accomplished. Previous approaches have focussed on how neural networks might produce PPCs from sensory input or perform calculations using them, like combining multiple PPCs. Less attention has been given to the question of how the necessary organisation of neurons can arise and how the required knowledge about the input statistics can be learned. In this paper, we propose a model of learning multi-sensory integration based on an unsupervised learning algorithm in which an artificial neural network learns the noise characteristics of each of its sources of input. Our algorithm borrows from the self-organising map the ability to learn latent-variable models of the input and extends it to learning to produce a PPC approximating a probability density function over the latent variable behind its (noisy) input. The neurons in our network are only required to perform simple calculations and we make few assumptions about input noise properties and tuning functions. We report on a neurorobotic experiment in which we apply our algorithm to multi-sensory integration in a humanoid robot to demonstrate its effectiveness and compare it to human multi-sensory integration on the behavioural level. We also show in simulations that our algorithm performs near-optimally under certain plausible conditions, and that it reproduces important aspects of natural multi-sensory integration on the neural level.     

ON THE STABILITY OF LAVES PHASES

ＯＮＴＨＥＳＴＡＢＩＬＩＴＹＯＦＬＡＶＥＳＰＨＡＳＥＳ￥Ｌｉ，Ｃｈｏｎｇｈｅ；Ｋａｎｇ，Ｄｅｓｈａｎ；Ｑｉｎ，Ｐｅｉ；Ｃｈｅｎ，Ｎｉａｎｙｉ（ＳｈａｎｇｈａｉＩｎｓｔｉｔｕｔｅｏｆＭｅｔａｌｌｕｒｇｙ，ＣｈｉｎｅｓｅＡｃａｄｅｍｙｏｆＳｃｉｅｎｃｅｓ...     

A Dynamic Neural Substrate and Automatic Perception Switching

This paper presents DENN, a dynamic neural network or neural substrate having a number of abilities that might allow it to play a useful role as a constituent of an artificial cognitive system, handling the task of low-level perceptual processing. DENN can adapt without supervision to new objects, is able to respond to patterns of activation from several objects presented simultaneously to it, and is able to automatically switch its perception between multiple objects. It is based on an ideal neural substrate as conjectured by Dimond (1980), having the twin capabilities of autonomous learning and memory, capabilities emerging due to the use of autonomous neurons. DENN has a pyramidal architecture and its neurons have topologically organized receptive fields. Through training, the neurons become feature detectors, with the higher level neurons responding to more complex features. The neurons respond to a retinal input with an oscillatory output whose frequency depends only on their own input. Due to developing phase differences, the higher level neurons can move out of phase relative to each other. Therefore, different inputs are recognized cyclically-a process we term 'automatic perception switching'. Experiments verified the system's ability of automatic perception switching, investigated its response to randomized images, and compared the performance of adaptive and non-adaptive versions of the neural substrate.     

基于人工神经网络的焊缝熔敷金属力学性能预测

基于神经网络预测技术提出了一种新思路。本文利用前向神经网络，使用BP算法对熔融金属的抗拉强度、屈服强度等力学性能指标进行研究。通过对实验数据样本进行训练，建立熔敷金属的抗拉强度、屈服强度与焊缝熔敷金属成分之间的隐性函数，用此神经网络，即可预测熔敷金属的抗拉强度和屈服强度。     

Evolution of Neural Controllers for Locomotion and Obstacle Avoidance in a Six-legged Robot

This article describes how the SGOCE paradigm has been used within the context of a 'minimal simulation' strategy to evolve neural networks controlling locomotion and obstacle avoidance in a six-legged robot. A standard genetic algorithm has been used to evolve developmental programs according to which recurrent networks of leaky-integrator neurons were grown in a user-provided developmental substrate and were connected to the robot's sensors and actuators. Specific grammars have been used to limit the complexity of the developmental programs and of the corresponding neural controllers. Such controllers were first evolved through simulation and then successfully downloaded on the real robot.     

Dynamic interaction of attractors across multiple cortical networks as a neural basis for intersensory facilitation

I propose a neural network model for intersensory facilitation and investigate its essential neuronal mechanisms. The model consists of sensory networks (SI, SII) and an integration network (IT). The integration network binds information derived from the sensory networks and sends feedback signals to them. Through a Hebbian learning process, point attractors representing individual features and objects are created in the dynamics of the sensory networks and integration network, respectively. The ongoing state of each network is a randomly itinerant state among these point attractors. When the SI network is stimulated with a feature (I_ n ) belonging to an object (O_ n ) with suprathreshold intensity, the point attractor corresponding to I_ n emerges in the SI network, but does not when stimulated with subthreshold intensity. Intersensory facilitation occurs when associate feature II_ n derived from the other modality belonging to the same object (O_ n ) is simultaneously presented to other sensory network SII, where the point attractor corresponding to feature I_ n emerges even when the sensory networks are stimulated with subthreshold intensity. I suggest here that the dynamic interaction of relevant point attractors across multiple neural networks is essential for intersensory facilitation, and that self-organized synaptic modulation effectively contributes to intersensory facilitation when crossmodal stimuli are separated in time.     

用BP算法建立ANN模型时学习率的选取

对于用BP算法学习的神经元网络,不同层中神经元进入饱和状态后对网络学习过程的危害程度是不同的。以3-3-1结构的网络及其用BP算法修正连接权值的过程,对比分析了输出层和隐含层神经元进入饱和后对网络学习过程的影响。并对实验证明了不同层采用不同学习率可以改善网络学习收敛速度。     

The use of the neural networks in the recognition of the austenitic steel types

The work shows the results obtained in recognition of different types of austenitic steels with an ultrasonic system that provides the necessary data towards two different neural networks. One of the neural networks (RNAU) used as input a vector containing processed data (propagation velocity and ultrasonic attenuation). The second neural network (AUFRAN) used the amplitude of digitized radio-frequency signal and its numerical Fourier transform as input vector.Two thirds of data obtained from three kinds of steels (W.1.4541, W.1.6903 and HP50) were used in the learning process. The last third of acquired data on these samples were used in the testing process. The obtained classification probabilities were above 98.3%. As a supplement, the testing process was extended to three other types of austenitic steels having different chemical compositions than those used in the learning process.     

High capacity,small world associative memory models

Models of associative memory usually have full connectivity or, if diluted, random symmetric connectivity. In contrast, biological neural systems have predominantly local, non-symmetric connectivity. Here we investigate sparse networks of threshold units, trained with the perceptron learning rule. The units are given position and are arranged in a ring. The connectivity graph varies between being local to random via a small world regime, with short path lengths between any two neurons. The connectivity may be symmetric or non-symmetric. The results show that it is the small world networks with non-symmetric weights and non-symmetric connectivity that perform best as associative memories. It is also shown that in highly dilute networks small world architectures will produce efficiently wired associative memories, which still exhibit good pattern completion abilities.     

Modelling 3-D cutting dynamics with neural networks

Identification of 3-D cutting dynamics requires an expensive experimental set-up and complicated analysis. Recently, time series methods were used to model cutting dynamics. This approach allows a simpler experimental set-uup and estimates the discrete transfer functions used for simulation and/or calculation of frequency domain characteristics of the system. In this paper, the use of neural networks is proposed to model the 3-D cutting dynamics. Neural networks can be trained using the same experimental set-up used for the time series methods. However, several time series models (for different cutting speeds) can be represented with a single neural network, and cutting forces can be studied for varying cutting speed conditions. Also, four neural networks were used to store the frequency domain characteristics of the thrust direction cutting force. In this study, the estimation errors for the neural networks were less than 7% of the defined range (the difference between the maximum and minimum of the data).     

Neural network solutions to the tool condition monitoring problem in metal cutting—A critical review of methods

A wide range of cutting tool monitoring techniques have been proposed and developed in the last decade, but only a few have found industrial applications, and a truly universally applicable system has still to be developed. In this paper a review of tool condition monitoring (TCM) systems, developed or implemented through application of neural networks, is provided. The review seeks to illustrate the extent of application of neural networks and the need for multiple source sensor signals in TCM systems. A critical analysis of methods is included and the trend in obtained results outlined.     

神经网络技术在土木结构健康监测中的应用

结构损伤诊断子系统是建立结构智能健康监测专家系统的核心问题。采用人工神经网络技术可以实现结构损伤的自动识别与定位，具有广阔的应用前景。介绍了基于人工神经网络的两级损伤识别策略，并对采用人工神经网络进行结构损伤诊断的网络输入参数与网络结构选择等关键问题进行了探讨。     

离心法制备梯度功能材料中内生颗粒的分布

采用人工神经网络研究了在不同型温、浇温和转速条件下以离心法制备Al-16%SiFGM时初晶硅的分布规律,并通过实验进行了验证,在建立神经网络模型时,以型温、浇温、转速等工艺参数作为人工神经网络的输入,以内生初晶硅分布的相对厚度作为输出,实验表明,预测结果与实际测定结果比较吻合,说明采用神经网络预测离心法制备梯度功能材料中内生颗粒的分布是可行的。