A class of algorithms is presented for training multilayer perceptrons using purely linear techniques. The methods are based upon linearization of the network using error surface analysis, followed by a contemporary l...
详细信息
A class of algorithms is presented for training multilayer perceptrons using purely linear techniques. The methods are based upon linearization of the network using error surface analysis, followed by a contemporary least-squares estimation procedure. Specific algorithms are presented for estimating weights node-wise and layer-wise and for estimating the entire set of network weights simultaneously. In several experimental studies, the node-wise method was superior to backpropagation and to an alternative linearization method due to M. Azimi-Sadjadi et al. (1990) in terms of number of convergences and convergence rate. The layer-wise and network-wise updating offer further improvement.< >
In this paper, an optimal multilevel stack filtering algorithm (the stack filters used at each level are designed to be optimal) is introduced. Optimal multilevel stack filter is capable of considering a larger window...
详细信息
The joint problem of target tracking and identification was discussed previously by the authors (1991), and a tracking/identification algorithm using an extended-Kalman-filter-based associative memory (EK-FAM) was dem...
详细信息
The joint problem of target tracking and identification was discussed previously by the authors (1991), and a tracking/identification algorithm using an extended-Kalman-filter-based associative memory (EK-FAM) was demonstrated through several examples. The convergence properties of the algorithm are discussed. Under the appropriate conditions, a contraction operator can be developed, using Banach space concepts that guarantee convergence of the algorithm.< >
＜正＞The joint problem of tracking and identification of a target from an airplane was discussed in[1].In this paper the problem is discussed using a different system structure which uses a neural *** assume that a s...
＜正＞The joint problem of tracking and identification of a target from an airplane was discussed in[1].In this paper the problem is discussed using a different system structure which uses a neural *** assume that a set of targets are distinguishable where indistinguishability between any two targets implies that their feature vectors are *** associative memory,which is implemented with an artificial neural network in a feedback loop,is developed as an *** the associative memory is presented with a feature vector it recalls the corresponding *** mathematical concepts in Hilbert space show that perfect identification is possible by using such an identifier in the noise free *** estimate is used to identify correct targets when targets are subject to certain conditions of noise corruption of the feature *** can combine this identifier with a modified extended Kalman filter to solve the combined target tracking and identification *** result of this technique is promising,and fast identification is *** demonstrate this technique through an example.
Stack filters are a class of sliding-window nonlinear digital filters that possess the weak superposition property (threshold decomposition) and the ordering property known as the stacking property. They have been dem...
详细信息
Stack filters are a class of sliding-window nonlinear digital filters that possess the weak superposition property (threshold decomposition) and the ordering property known as the stacking property. They have been demonstrated to be robust in suppressing noise. Two methods are introduced to adaptively configure a stack filter. One is by employing the least mean square (LMS) algorithm and the other is based on perceptron learning. Experimental results are presented to demonstrate the effectiveness of the methods for noise suppression.< >
The generalized adaptive neural filter (GANF), a new class of nonlinear filters, is introduced. It is effective for non-Gaussian noise suppression. Some properties of GANF are derived, and an algorithm for finding the...
详细信息
The generalized adaptive neural filter (GANF), a new class of nonlinear filters, is introduced. It is effective for non-Gaussian noise suppression. Some properties of GANF are derived, and an algorithm for finding the optimal GANF, based on the upper bound in the minimum absolute error, is proposed. The implementation of the optimal GANF by using the least mean square error and the least perceptron error is also discussed. Experimental results are presented to demonstrate the effectiveness of the new filter.< >
The communication latency problem is presented with special emphasis on RISC (reduced instruction set computer) based multiprocessors. An interprocessor communication model for parallel programs based on locality is p...
详细信息
The communication latency problem is presented with special emphasis on RISC (reduced instruction set computer) based multiprocessors. An interprocessor communication model for parallel programs based on locality is presented. This model enables the programmer to manipulate locality at the language level and to take advantage of currently available system hardware to reduce latency. A hardware node architecture for a latency-tolerant RISC-based multiprocessor, called Seamless, that supports this model, is presented. The Seamless architecture includes the addition of a hardware locality manager to each processing element, as well as an integral runtime environment and compiler.< >
The optical properties of thin film microcrystalline silicon (μc-Si:H) prepared by plasma-enhanced chemical vapor deposition (PECVD) have been studied by real time spectroscopie ellipsometry in the nucleation regime ...
The optical properties of thin film microcrystalline silicon (μc-Si:H) prepared by plasma-enhanced chemical vapor deposition (PECVD) have been studied by real time spectroscopie ellipsometry in the nucleation regime as isolated crystalline particles increase in size. A simple geometric model of nucleation allows us to remove the dominant effect of voids and extract the dielectric functions of the crystallites themselves. We find that the results can be understood in terms of a classical size effect whereby limitations on the electron mean free path by scattering at crystallite surfaces control the absorption onset from 2.0 to 3.0 eV. Finally, we describe how well-ordered, continuous 15 Å c-Si films can be prepared on metal substrates.
This paper focuses on the upper bounds for both the mean delay and the probability of cell loss that bursty arrivals incur in a finite capacity multiqueue system with nonexhaustive cyclic service. We compute the upper...
详细信息
This paper focuses on the upper bounds for both the mean delay and the probability of cell loss that bursty arrivals incur in a finite capacity multiqueue system with nonexhaustive cyclic service. We compute the upper bounds for this system by considering a cell multiplexer with the same arrival processes and equal queue capacity. Under the ATM environment, the mean delay obtained from this multiplexer cannot only serve as an upper bound but also render a fairly accurate estimation for the mean delay of the polling system. For the cell loss probability, we consider a multiple urn model with uniform occupancy distribution which will guarantee the upper bound. A heuristic method is proposed to give better estimates for cases which have medium to high cell loss rate.
A mathematical framework for the solution of statistical inference problems on a class of random sets is proposed. It is based on a new definition of expected pattern. The least-mean-difference estimator (restoration ...
详细信息
A mathematical framework for the solution of statistical inference problems on a class of random sets is proposed. It is based on a new definition of expected pattern. The least-mean-difference estimator (restoration filter) is proved, under certain conditions, to be equivalent to the minimization of the measure of size (area) of the set-difference between the original pattern and the expected pattern of the estimated (restored) pattern. Consequently, it is proved that, under certain conditions, if the estimator (restoration filter) is unbiased, then it is the least mean difference estimator (restoration filter).< >
暂无评论