最新刊期
卷 ， 期 10 ， 1992
- 摘要：Our main goal is to explore the application of neural networks to minimum weighted distance decoding. At the beginning, we review some models and convergence properties of neural networks step by step. Next, we study properties and applications of the minimum weighted distance decoding. Finally, we prove that finding the global minimum point of any energy function is equivalent to minimum weighted distance decoding in some （n, k） linear code, and propose an algorithm which uses annealing method with Boltzmann machine for minimum weighted distance decoding.关键词：neural network;Minimum weighted distance decoding;Boltzmann machime;Annealing2|73|3更新时间：2025-12-08
- 摘要：<正> 由中国电子学会、美国佐治亚理工学院等单位共同主办的第二届国际毫米波与远红外技术会议（ICMWFT,92）于1992年8月17日至21日在北京召开,来自澳大利亚、比利时、加拿大、法国、日本、埃及、瑞士、乌克兰、俄罗斯、美国、英国等11个国家的31名代表与中国的75名论文作者参加了会议. 会议主席M.Von Orfenberg,程序委员会中方主席胡汉泉,美方主席R.W.McM:llan在开幕式上分别致词.为纪念此系列会议的创始人之一、上届会议主席J.J.Gallagher博士,McMillan主席还在开幕式上简2|11|0更新时间：2025-12-08
- 摘要：Instead of using statistical nature techniques, which are popular in neural network analysis, straightforward mathematical analysis is developed in this paper to promote understanding of properties of the Hopfield network when it is used as a CAM. By introducing a concept of non-orthogonal degree d of a p-set n-binary stored patterns, the capacity of the network, i. e., the maximum number of stable stored patterns, is （n + d）/（d + 1） in the worst case; and the stability degree k*, a quantity describing the radius of convergence basin, is positively proportional to n - p - （p - 1） d and inversely proportional to p. Concepts of associative pattern and spurious pattern are considered; About spurious patterns, we have provedthat in some special cases sgn （） are spurious patterns, where关键词：Hopfield network;capacity;Associative patterns;Spurious patterns2|53|0更新时间：2025-12-08
- 摘要：In this paper, we propose an Equal-Error Range Approximation and Shrinking Learning Algorithm for multilayer perceptrons. It requires the error between each network output node activation and its target to fall into a given error range, thus it can learn faster in lower calculation cost and may avoid reversed target output and overlearning. hence itcan improve networks’generalization abilities in pattern recognitions. Through gradually Shrinking of the error range, it can also enable the networks to learn the targets more accurately in less training iterations. Finally, we apply this learning algorithm trained network to the EEG detection, and the experiment results have showed the above advantages of the proposed algorithm.关键词：neural networks;Equal-error range;pattern recognition2|42|3更新时间：2025-12-08
- 摘要：A new neural network model-general-purpose master-slave neural] networkmodel is presented in this paper, and two master-slave control methods are discussed which prove the property of general purpose of this model.关键词：Master-slave neural network;energy function;Local minimum;System stab-ability2|62|2更新时间：2025-12-08
- 摘要：This paper presents a 2-D neural network model with the applications in solving a class of matrix equations. We show both by analysis and by simulations that this network is guaranteed to provide the results arbitrarily close to the accurate ones within a few of the time constants of this network.关键词：neural network;Matrix equation;optimization2|63|2更新时间：2025-12-08
- 摘要：Level by Level Learning for Artificial Neuronal Groups is proposed in this paper. The learning process suggested by the method is more similar to the process of knowledge growth for human individual and society, and can improve generalization ability and learning efficiency for network. This paper carries out some case studies.关键词：Artificial neuronal groups;Level by level learning2|39|2更新时间：2025-12-08
- 摘要：A new neural network for linear programming is proposed in this paper. It is shown both analytically and by simulations that our network is guaranteed to be asymptotically stable in the large and to provide the results arbitrarily close to the optimal solutions of linear programming problems.关键词：neural network;linear programming;optimization2|121|4更新时间：2025-12-08
- 摘要：A neural network with floating gate transistors has been designed. It has the features of continuously changed weights, simple structure, small size units, etc. These chips can be cascaded to form a large network due to characteristics of distributed neuron structure. Therefore, this network is feasible for various applications. 8×8 full interconnected neural network chip has been fabricated in 3μm floating gate NMOS process. It is composed of 128 programmable floatimg gate NMOS transistors, which corresponds to full interconnected network with 8 neurons. The specific applications for the chip has been studied as digit recognition and binary image processing. The results demonstrate that it has prospects of actual applications and large flexibility. In addition, it is easy to be fabricated since the structure is simple and suitable for IC process.关键词：artificial neural network;IC;Circuits Principle and design2|113|2更新时间：2025-12-08
- 摘要：A neural network modelling method for nonlinear plants by using a new learning algorithm, Generalized Self-organized Learning （GSL）, is proposed in this paper. This method employs multiple local models for plant modelling. It develops the division mechanism. for local models adopted in the Kohonen self-organized learning algorithm, and the achieved local models take both the distribution of input samples and the model matching errors into account. Simulation results show that the GSL algorithm improves modelling accuracy and learning speed obviously.关键词：neural network;Nonlinear plant modelling;Self-organized learning;Back propagation learning;Pattren clustering2|44|3更新时间：2025-12-08
- 摘要：The optimum learning rate BP algorithm is reviewed for addressing problem-involved in numerical experiments and formulas to efficiently implement it when applied to several net architectures most in use. The simulation results further release properties of the algorithm.关键词：Optimum learning rate;BP algorithm;Multilayer neural networks;Predic-t ion2|102|2更新时间：2025-12-08
- 摘要：The optical implementation of WTA neural network model with bipolar input-layer neurons and bipolar input-layer-to-hidden-layer connections is presented. The input to the hidden-layer neurons is the bipolarly weighted summation of the bipolar neuron states of the input-layer, which is equivalent to the unipolarly weighted summation of the unipolar input states and the summation of the unipolar reversed storage pattern by a constant 1/2. The constant 1/2 can be implemented by dividing each neural pixel into two equal parts of which one is opaque and the other is transparent. Experimental results shows that both the storage capacity and the content addressability of the optical system are improved compared to those of the unipolar WTA optical system关键词：WTA neural nnetwork model;Bioplar neural state;pattern recognition;Multichannel inner-product hologram.2|42|3更新时间：2025-12-08
- 摘要：In this work, a neural network approach [for the classification of aircrafts under affine transformation is described. The method of Fourier transformation has been extended to produce a set of normalized invariants which are independent of the affine transformation and the starting point. A threelayer perceptron is trained with these invariants using backpropagation. By adopting an accelerated learning algorithm, the learning time can be reduced greatly. Typical results of the classification of aircrafts and a good performance of noise tolerance are presented.关键词：affine transformation;invariants;neural network;backpropagation;Classification2|31|0更新时间：2025-12-08
- 摘要：<正> 作为现代社会政治、经济,文化与军事活动重要支柱之一的通信系统,已成为当今信息与电子产业发展的重点.随着计算机与微电子技术的飞速发展,现代通信系统与设备中普遍采用软件技术与专用超大规模集成电路技术.专家认为:没有通信软2|21|0更新时间：2025-12-08
- 摘要：This paper reports an implementation of dynamic programming based time-normalized algorithm, called dynamic time warping （DTW）, on neural networks. DTW is one of the most successful algorithms for spoken word recognition. It is very robust and usually provides the highest recognition rate possible but it takes a lot of computer time unless it is implemented by special hardware. In this implementation, the computation is governed by two recurrent subnets and one memory layer, demonstrating a hard-wiring mechanism which benefits from existing approaches.关键词：neural networks;speech processing;pattern recognition2|118|1更新时间：2025-12-08
- 摘要：The article presents a logic system called as approximate logic system, whi-ch has several coefficients. The fuzzy logical value will be allowed to be used in the logic system and the logic operator can also be fuzzy. Finally, a neural system has been developed based on approximate logic system and applied to a multi-experts’ opinion synthesis system.关键词：Approximate logic;And-or degree of operator;Expert opinion synthesis2|19|5更新时间：2025-12-08
- 摘要：Some properties of neural optimization are studied in this paper. The properties of the solutions and the region of attraction, the division of the feasible solution space are investigated.关键词：neural networks;optimization;algorithm;Stable points;Attractive area2|55|3更新时间：2025-12-08
- 摘要：Kohonen’s self-organizing feature maps neural network was applied to solve the quadratic assignment problem, a special case of VLSI placement when all nets have two terminals, Compared with the Min-Cut algorithm, neural placement algorithm can obtain optimal results in less time for small problems. Computing time for good placement was lengthened as problem size grew.关键词：Quadratic assignment;placement;artificial neural networks2|33|6更新时间：2025-12-08
- 摘要：This paper gives a review on the recent advances on supervised learnieg methods for static feedforward networks, i.e （1） Various improvements and variants made on the classical Back Propagation. （2） A number of other learning methods for training Multilayer Perceptron. （3） Other feedforward models and supervised learning models. （4） Models with complex structures.关键词：Feedforward networks;Supenrised learning;neural networks2|78|9更新时间：2025-12-08
- 摘要：An attempt is made in the paper to give an exposition on artificial neural networks （ANNs） concerning its mechanisms, strength and weakness as well as future trends and strategies that may interest the ANN researchers.4|167|3更新时间：2025-12-08
- 摘要：<正> 中国神经网络委员会CNNC（China Neural Network Council）于1992年8月24日在北京邮电学院召开了全会.会上研究了IJCNN’92-Beijing的进展情况和下一步的工作安排. IJCNN（International Joint Conference on Neural Networks）是由国际神经网络学会INNS（Internation-al Neural Network Society）和IEEE神经网络委员会IEEE NNC(Institute of Electrical and Electronics2|16|0更新时间：2025-12-08
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0