具有期望容错域的前向掩蔽联想记忆模型的设计方法

联想记忆的综合问题是目前没有很好解决的难题.文中用作者提出的通用前馈网络和排序学习算法,提出了一种设计具有期望容错域的前向掩蔽联想记忆模型的方法.该方法一般性地解决了信息空间上联想记忆的综合难题,使设计出的联想记忆模型具有任意期望的记忆样本容错域.     

容量约束的自组织增量联想记忆模型

自组织联想记忆神经网络因其并行、容错及自我学习等优点而得到广泛应用,但现有主流模型在增量学习较大规模样本时,网络节点数可能无限增长,从而给实际应用带来不可承受的内存及计算开销。针对该问题,提出了一种容量约束的自组织增量联想记忆模型。以网络节点数为先决控制参数,结合设计新的节点间自竞争学习策略,新模型可满足大规模样本的增量式学习需求,并能以较低的计算容量取得较高的联想记忆性能。理论分析表明了新模型的正确性与有效性,实验分析同时显示了新模型可有效控制计算容量,提升增量样本学习效率,并获得较高的联想记忆性能,从而能更好地满足现实应用需求。     

基于小世界体系的指数自联想记忆模型研究

全互连结构是大多数神经联想记忆模型采用的连接方式．然而,无论从神经生物学的观点还是从硬件实现角度出发,在确保局部连接占主体地位的情况下,最小化神经元间的连接长度是合乎逻辑的．由Watts和Strogatz提出的小世界体系（SWA）为解决这一问题提供了一条有趣途径．借鉴这一具有一定神经生物学特点的小世界体系,构建出一个基于小世界体系的指数自联想记忆模型（SWEAM）,其在结构上更接近人脑的生物学特性,同时在存储容量和抗噪声性能上优于现有的同类联想记忆模型．因采用了SWA,SWEAM在工艺实现上也更容易．     

联想记忆神经网络的一个有效学习算法

提出一种新的联想记忆网络模型的有效学习算法,它具有下述特点:(1)可以全部存 储任意给定的训练模式集,即对于训练模式的数目和它们之间相关性的强弱没有限制;(2)最 小的训练模型吸引域达到最大;(3)在(2)的基础上,每个训练模式具有尽可能大的吸引域; (4)联想记忆神经网络是全局稳定的.大量的计算机仿真实验结果充分说明所提出的学习算 法比已有算法具有更强的存储能力和联想容错能力.     

GAM模型稳定性分析及其在图像识别中的应用

多值指数关联联想记忆模型（MMECAM）是一种高存储容量的自联想记忆神经网络。在详细分析其优缺点的基础上,通过改进MMECAM模型的更新规则,首先提出一个新的高斯自联想记忆模型（GAM）,然后通过定义简单的能量函数从理论上证明其在同、异步方式下的稳定性,从而保证所存储的模式能最终成为GAM的稳定点;其次,通过引入一般相似性测度进一步提出广义GAM模型（G-GAMs）框架,使得GAM模型成为其特例;最后,将GAM模型应用于单样本图像识别,计算机模拟证实了该模型的鲁棒性能。     

Hopfield网络联想记忆外积法设计权的研究

首先验证了离散Hopfield网络联想记忆外积法设计权的方法，然后通过公式逐步推导，得出了一个特殊矩阵。通过计算机仿真模拟验证了样本容错能力与样本相关性的关系，从输入图形和输出图形来看，联想效果尚可，具有一定的实效性。但也存在着一些问题，比如有些模式不能完全记忆。     

医疗诊断专家系统推理机的AM网络设计

本文探讨了医疗诊断专家系统（ＭＤＥＳ）的特殊性，提出了适合于ＭＤＥＳ推理机的联想记忆神经网络（ＡＭＮＮ）。利用该网络模型不要求样本独立和二值输入等特性，引入隶属度，将模糊模式识别运用于ＭＤＥＳ《     

协同系统中可扩展的动态容错模型研究及实现

针对协同系统的特点以及其对可靠性、实时性和性能等方面的要求，提出了一种可扩展的动态容错模型xDFT（Extensible Dynamic Fault Tolerance Model，xDFT）．本容错模型将容错支持结构和容错机制结合在一起：在本模型中首先定义了一种可扩展的、动态容错支撑结构；基于此，提出了一种容错机制．xDFT模型能够根据系统负载状况设定服务节点的负载阈值，改变服务冗余度，不仅提高了服务效率，而且以一种简单有效的方式实现了负载平衡．     

一种基于稀疏分布记忆模型的汉字联想记忆方法

本文描述了Ｋａｎｅｒｖａ的稀疏分布记忆模型，指出了它在用于汉字联想时的问题，同时提出了改进的模型，试验表明，这种改进模型使记忆容量和容错能力大大提高。     

讲座：人工神经网络：第四讲 反馈网络与联想记忆（续）

4基本学习规则Hopfield网络的学习，采用的是无导师学习。学习的过程相应于形成网络的连接权矩阵W。讨论的中心问题在于，如何使网络对于给定的问题进行学习，建立连接权矩阵，并使网络具有较强的联想记忆能力。假设需要存贮的记忆样本有P个X1，X2，…，XPXi∈{－1， 1}n（1）Hebb学习规则a．对一个模式的学习为了分析简便，首先考虑网络对一个模式的学习。这时需要存贮的记忆样本只有1个（设为X1），它将成为网络的稳定状态，并具有最大的吸引域，或者说具有最大的“纠错能力”。将记忆样本X1输入到网络，并作为网络的初始状态，经过…     

基于模式关联的一对多联想记忆方法

在现有的多模块一对多联想记忆模型中,由于所处理的记忆模式集合本身的特点以及记忆模式之间的关联被忽视,使得构造出来的模型结构复杂,难以实际应用.针对这一不足,提出一种基于模式关联的实现方法.以该方法构造出的多模块一对多联想记忆模型结构简单,易于硬件实现,使得多模块一对多联想记忆模型具有了实际应用的可能.     

A columnar model explaining long-term memory

A hologram provides a useful model for explaining the associative memory of the brain. Recent advances in neuroscience emphasize that single neurons can store complex information and that subtle changes in neurons underlie the process of memorization. Experimental results suggest that memory has a localized character. This finding is inconsistent with the characteristics of holographic memory, because this memory has a delocalized, uniform distribution of refractive index in the recorded medium. The recently proposed columnar memory model has a discrete distribution of refractive index. In this study, we first examined the performance of columnar memory and found that it was comparable to holographic memory. Secondly, we showed that this model could explain the above-mentioned experimental results as well as associative memory.     

Face recognition by applying wavelet subband representation and kernel associative memory

In this paper, we propose an efficient face recognition scheme which has two features: 1) representation of face images by two-dimensional (2D) wavelet subband coefficients and 2) recognition by a modular, personalised classification method based on kernel associative memory models. Compared to PCA projections and low resolution "thumb-nail" image representations, wavelet subband coefficients can efficiently capture substantial facial features while keeping computational complexity low. As there are usually very limited samples, we constructed an associative memory (AM) model for each person and proposed to improve the performance of AM models by kernel methods. Specifically, we first applied kernel transforms to each possible training pair of faces sample and then mapped the high-dimensional feature space back to input space. Our scheme using modular autoassociative memory for face recognition is inspired by the same motivation as using autoencoders for optical character recognition (OCR), for which the advantages has been proven. By associative memory, all the prototypical faces of one particular person are used to reconstruct themselves and the reconstruction error for a probe face image is used to decide if the probe face is from the corresponding person. We carried out extensive experiments on three standard face recognition datasets, the FERET data, the XM2VTS data, and the ORL data. Detailed comparisons with earlier published results are provided and our proposed scheme offers better recognition accuracy on all of the face datasets.     

训练模式摄动对模糊形态学联想记忆网络的影响

众多学者研究的两类形态学联想记忆网络的存储能力、抗腐蚀/膨胀噪声的能力等性质几乎都相同。但是文中研究发现两类网络对训练模式摄动的鲁棒性差异很大。一类对训练模式摄动拥有好的鲁棒性,而另一类则较差。该研究结论能为形态学联想记忆网络的学习算法选择和训练模式采集设备的精度要求提供指导,对前期训练模式的获取过程提供警示。     

State space model predictive fault-tolerant control for batch processes with partial actuator failure

This paper presents a state space model predictive fault-tolerant control scheme for batch processes with unknown disturbances and partial actuator faults. To develop the model predictive fault-tolerant control, the batch process is first treated into a non-minimal representation using state space transformation. The relevant concepts of the corresponding model predictive fault-tolerant control is thus introduced through state space formulation, where improved closed-loop control performance is achieved even with unknown disturbances and actuator faults, because, unlike traditional model predictive fault-tolerant control, the proposed control method can directly regulate the process output/input changes in the design. For performance comparison, a traditional model predictive fault-tolerant control is also designed. Application to injection velocity control shows that the proposed scheme achieve the design objective well with performance improvement.     

Stability analysis of bidirectional associative memory networks with time delays

By using the method of Liapunov functional, a model for bidirectional associative memory networks with time delays is studied. The asymptotic stability is global in the state space of the neuronal activations and is also independent of the delays. Our results can be applied to a variety of situations that arise both in the field of biological and artificial neural networks.     

基于结构学习和迭代自映射的自联想记忆模型

传统的人工神经元网络连接结构是固定的,是对权值的学习.提出一种基于生理神经元特征的人工神经元模型,并在以此为单元构成的用于实现自联想记忆的神经网络上进行对结构的学习.学习算法以设定神经元的输入/输出感受野、调整突触和轴突末梢的连接、并行的自投影迭代为特征.给出了此网络模型的矩阵描述和实验结果.     

Part 3: Brain science, information science and associative memory model

I review recent progress on the associative memory model, which is a kind of neural network model. First, I introduce this model and a mathematical theory called statistical neurodynamics describing its properties. Next, I discuss an associative memory model with hierarchically correlated memory patterns. Initially, in this model, the state approaches a mixed state that is a superposition of memory patterns. After that, it diverges from the mixed state, and finally converges to a memory pattern. I show that this retrieval dynamics can qualitatively replicate the temporal dynamics of face-responsive neurons in the inferior temporal cortex, which is considered to be the final stage of visual perception in the brain. Finally, I show an unexpected link between associative memory and mobile phones (CDMA). The mathematical structure of the CDMA multi-user detection problem resembles that of the associative memory model. It enables us to apply a theoretical framework of the associative memory model to CDMA.     

Auto-associative memory with two-stage dynamics of nonmonotonicneurons

Dynamical properties of a neural auto-associative memory with two-stage neurons are investigated theoretically. The two-stage neuron is a model whose output is determined by a two-stage nonlinear function of the internal field of the neuron (internal field is a weighted sum of outputs of the other neurons). The model is general, including nonmonotonic neurons as well as monotonic ones. Recent studies on associative memory revealed superiority of nonmonotonic neurons to monotonic ones. The present paper supplies theoretical verification on the high performance of nonmonotonic neurons and proves that the capacity of the auto-associative memory with two-stage neurons is O(n/ radicallog n), in contrast to O(n/log n) of simple threshold neurons. There is also a discussion of recall processes, where the radius of basin of attraction of memorized patterns is clarified. An intuitive explanation on why the performance is improved by nonmonotonic neurons is also provided by showing the correspondence of the recall processes of the two-stage-neuron net and orthogonal learning.