A genetic method to generate a neural network which has both structure and connection weights adequate for a given task is proposed. A neural network having arbitrary connections is regarded as a virtual living thing which has genes representing its connections among neural units. Effectiveness of the network is estimated from its time sequential input and output signals. Excellent individuals, namely appropriate neural networks, are generated through generation iterations. The basic principle of the method and its applications are described. As an example of evolution from randomly generated networks to feedforward networks, an XOR problem is dealt with, and an action control problem is used for making networks containing feedback and mutual connections. The proposed method is available for designing a neural network whose adequate structure is unknown.

The method of estimating multiple sound source locations based on a neural network algorithm and its performance are described in this paper. An evaluation function is first defined to reflect both properties of sound propagation of spherical wave front and the uniqueness of solution. A neural network is then composed to satisfy the conditions for the above evaluation function. Locations of multiple sources are given as exciting neurons. The proposed method is evaluated and compared with the deterministic method based on the Hyperbolic Method for the case of 8 sources on a square plane of 200m200m. It is found that the solutions are obtained correctly without any pseudo or dropped-out solutions. The proposed method is also applied to another case in which 54 sound sources are composed of 9 sound groups, each of which contains 6 sound sources. The proposed method is found to be effective and sufficient for practical application.

Abrupt variations of attractors caused by argumental discreteness in non-Hermitian complex-valued neural networks are reported. When we apply the complex-valued associative memories to dynamical processing, the weighting matrices are constructed as non-Hermitian in general so that they have motive force to the signal vectors. It is observed that competitions between argumental rotation force and noise-suppression ability of associative memories lead to trajectory distortions and abrupt variations of the attractors.

Asynchronous machines are a topic of great interest in the research area of actuators. Due to the complexity of these systems and to the required performance, the modelling and control of asynchronous machines are complex questions. Problems arise when the control goals require accurate descriptions of the electric machine or when we need to identify some electrical parameters; in the models employed it becomes very hard to take into account all the phenomena involved and therefore to make the error amplitude adequately small. Moreover, it is well known that, though an efficient control strategy requires knowledge of the flux vector, direct measurement of this quantity, using ad hoc transducers, does not represent a suitable approach, because it results in expensive machines. It is therefore necessary to perform an estimation of this vector, based on adequate dynamic non-linear models. Several different strategies have been proposed in literature to solve the items in a suitable manner. In this paper the authors propose a neural approach both to derive NARMAX models for asynchronous machines and to design non-linear observers: the need to use complex models that may be inefficient for control aims is therefore avoided. The results obtained with the strategy proposed were compared with simulations obtained by considering a classical fifth-order non-linear model.

In this article a Neural Network learning scheme is described, which is a generalization of VQ (Vector Quantization) and ART2a (a simplified version of Adaptive Resonance Theory 2). The basic differences between VQ and ART2a will be exhibited and it will be shown how these differences are covered by the generalized scheme. The generalized scheme enables a rich set of variations on VQ and ART2a. One such variation uses the expression ||I||2+||zj||2/||zj||sin(I,zj), as the distance measure between input vector I and weight vector zj. This variation tends to be more robust to noise than ART2a, as is shown by experiments we performed. These experiments use the same data-set as the ART2a experiments in Ref.(3).

A new optimal nonlinear regulator design method is developed by applying a multi-layered neural network and a fixed point theorem for a nonlinear controlled system. Based on the calculus of variations and the fixed point theorem, an optimal control law containing a nonlinear mapping of the state can be derived. Because the neural network has not only a good learning ability but also an excellent nonlinear mapping ability, the neural network is used to represent the state nonlinear mapping after some learning operations and an optimal nonlinear regulator may be formed. Simulation demonstrates that the new nonlinear regulator is quite efficient and has a good robust performance as well.

The three layer neural network (TLNN) is treated, where the nonlinearity of a neuron is of signum. First we propose an expression of the discriminant function of the TLNN, which is called a linear-homogeneous expression. This expression allows the differentiation in spite of the signum property of the neuron. Subsequently a learning algorithm is proposed based on the linear-homogeneous form. The algorithm is an error-correction procedure, which gives a mathematical foundation to heuristic error-correction learnings described in various literatures.

A model of a single neuron with chaotic dynamics is implemented with current-mode circuit design technique. The existence of chaotic dynamics in the circuit is demonstrated by simulation with SPICE3. The proposed circuit is suitable for implementing a chaotic neural network composed of such neuron models on a VLSI chip.

In this work, we introduce a fuzzy inference in conventional backpropagation learning algorithm, for networks of neuron like units. This procedure repeatedly adjusts the learning parameters and leads the system to converge at the earliest possible time. This technique is appropriate in a sense that optimum learning parameters are being applied in every learning cycle automatically, whereas the conventional backpropagation doesn't contain any well-defined rule regarding the proper determination of the value of learning parameters.

This paper describes a text-independent speaker recognition method using predictive neural networks. For text-independent speaker recognition, an ergodic model which allows transitions to any other state, including selftransitions, is adopted as the speaker model and one predictive neural network is assigned to each state. The proposed method was compared to quantization distortion based methods, HMM based methods, and a discriminative neural network based method through text-independent speaker identification experiments on 24 female speakers. The proposed method gave the highest identification rate of 100.0%, and the effectiveness of predictive neural networks for representing speaker individuality was clarified.

This article analyzes the property of the fully interconnected neural networks as a method of solving combinatorial optimization problems in general. In particular, in order to escape local minimums in this model, we analyze theoretically the relation between the diagonal elements of the connection matrix and the stability of the networks. It is shown that the position of the global minimum point of the energy function on the hyper sphere in n dimensional space is given by the eigen vector corresponding the maximum eigen value of the connection matrix. Then it is shown that the diagonal elements of the connection matrix can be improved without loss of generality. The equilibrium points of the improved networks are classified according to their properties, and their stability is investigated. In order to show that the change of the diagonal elements improves the potential for the global minimum search, computer simulations are carried out by using the theoretical values. In according to the simulation result on 10 neurons, the success rate to get the optimum solution is 97.5%. The result shows that the improvement of the diagonal elements has potential for minimum search.

An adaptive algorithm is presented for fuzzy clustering of data. Partitioning is fuzzified by addition of an entropy term to objective functions. The proposed method produces more convex membership functions than those given by the fuzzy c-means algorithm.

This paper discusses multiple-valued memory circuit using floating gate devices. It is an object of the paper to provide a new and improved analog memory device, which permits the memory of an amount of charges that accurately corresponds to analog information to be stored.

In this paper, two models for associative memory based on a measure of manhattan length are proposed. First, we propose the two-layered model which has an advantage to its implementation by using PDN. We also refer to the way to improve the recalling ability of this model against noisy input patterns. Secondly, we propose the other model which always recalls the nearest memory pattern in a measure of manhattan length by lateral inhibition. Even if a noise of input pattern is so large that the first model can not recall, this model can recall correctly against such a noisy pattern. We also confirm the performance of the two models by computer simulations.

In general, the work on developing an expert system has relied on domain experts to provide all domain-specific knowledge. The method for acquiring knowledge directly from experts is inadequate in oriental medicine because it is hard to find an appropriate expert and the development cost becomes too high. Therefore, we have developed two effective methods for acquiring knowledge indirectly from sample cases. One is to refine a constructed knowledge base by using sample cases. The other is to train a neural network by using sample cases. To demonstrate the effectiveness of our methods, we have implemented two prototype systems; the Oriental Medicine Expert System (OMES) and the Oriental Medicine Neural Network (OMNN). These systems have been compared with the system with the knowledge base built directly by domain experts (OLDS). Among these systems, OMES are considered to be superior to other systems in terms of performances, development costs, and practicalness. In this paper, we present our methods, and describe our experimental and comparison results.

This letter describes the concepts that the learnability of multilayer neural networks exists in a constrained hypersurface in learning space which is formed by input and output subspace of multilayer neural networks, and that a priori information, providing constraints on the learning space, is required for generalization.

An information retrieval system based on a dynamic thesaurus was developed utilizing the connectionist approach. The dynamic thesaurus consists of nodes, which represent each term of a thesaurus, and links, which represent the connections between nodes. Term information that is automatically extracted from user's relevant documents is used to change node weights and generate links. Thus, node weights and links reflect a user's interest. A document retrieval experiment using the dynamic thesaurus was conducted in which both a high recall rate and a high precision rate were achieved.

This article discusses a synthesis procedure of a discrete-time asynchronous neural network whose information is a limit cycle. The synthesis procedure uses a novel connection matrix and can be reduced into a linear epuation. If all elements of desired limit cycles are independent at each transition step, the equation can be solved and all desired limit cycles can be stored. In some experiments, our procedure exhibits much better storing performance than previous ones.

This paper describes a VLSI processor architecture designed for a back-propagation accelerator. Three techniques are used to accelerate the simulation. The first is a multi-processor approach where a neural network simulation is suitable for parallel processing. By constructing a ring network using several processors, the simulation speed is multiplied by the number of the processors. The second technique is internal parallel processing. Each processor contains 4 multipliers and 4 ALUs that all work in parallel. The third technique is pipelining. The connections of eight functional units change according to the current stage of the back-propagation algorithm. Intermediate data is sent from one functional unit to another without being stored in extra registers and data is processed in a pipeline manner. The data is in 24-bit floating point format (18-bit mantissa and 6-bit oxponent). The chip has about 88,000 gates, including microcode ROM for processor control, the processor is designed using 0.8-µm CMOS gate arrays, and the estimated performance at 40 MHz is 20 million connection updates per second (MCUPS). For a ring network with 4 processors, performance can be enhanced up to 90 MCUPS.

A novel harmonic retrieval algorithm is proposed in this paper based on Hopfield's neural network. Frequencies can be retrieved with high accuracy and high resolution under low signal to noise ratio (SNR). Amplitudes and phases in harmonic signals can also be estimated roughly by an energy constrained linear projection approach as proposed in the algorithm. Only no less than 2q neurons are necessary in order to detect harmonic siglnals with q different frequencies, where q denotes the number of different frequencies in harmonic signals. Experimental simulations show fast convergence and stable solution in spite of low signal to noise ratio can be obtained using the proposed algorithm.