Showing papers on "Support vector machine published in 1989"
TL;DR: The authors describe several fundamentally useful primitive operations and routines and illustrate their usefulness in a wide range of familiar version processes in terms of a vector machine model of parallel computation.
Abstract: The authors describe several fundamentally useful primitive operations and routines and illustrate their usefulness in a wide range of familiar version processes. These operations are described in terms of a vector machine model of parallel computation. They use a parallel vector model because vector models can be mapped onto a wide range of architectures. They also describe implementing these primitives on a particular fine-grained machine, the connection machine. It is found that these primitives are applicable in a variety of vision tasks. Grid permutations are useful in many early vision algorithms, such as Gaussian convolution, edge detection, motion, and stereo computation. Scan primitives facilitate simple, efficient solutions of many problems in middle- and high-level vision. Pointer jumping, using permutation operations, permits construction of extended image structures in logarithmic time. Methods such as outer products, which rely on a variety of primitives, play an important role of many high-level algorithms. >
89 citations
1 Mar 1989
TL;DR: The two most common algorithms for computing nonlinear control problems, as used in economics, are investigated on their vectorizability.
Abstract: In general, the econometric models relevant for policy evaluation contain a large number of equations. Therefore, in applying optimal control techniques, computational difficulties are encountered. In this paper the two most common algorithms for computing nonlinear control problems, as used in economics, are investigated on their vectorizability.
15 citations
TL;DR: A new version of the frontal technique which is designed for CRAY-like computers and is well suited to vector processing is described, effective in improving the efficiency of finite element solutions on a vector machine.
Abstract: At the kernel of most scientific computations lies the solution of linear equations which, in finite element codes, is often performed by the frontal method. This method, like many algorithms for sparse matrices, is usually implemented with extensive use of indirect addressing which scarcely benefits from vectorization. As a consequence, present application codes only partially exploit the innovative features of supercomputer architectures. This paper describes a new version of the frontal technique which is designed for CRAY-like computers and is well suited to vector processing. Numerical tests have shown that the proposed algorithm is effective in improving the efficiency of finite element solutions on a vector machine.
14 citations
1 Jan 1989
TL;DR: Experimental results indicate that the SVM can perform as well as the BP model, and the experiment indicates that distributed or local representation is developed by the learning algorithm.
Abstract: Summary form only given. The application of a stochastic vector machine (SVM) to alphanumeric character recognition is considered. The SVM is a new multilayered network with learning ability as in the backpropagation (BP) model. The system dynamics in the network is represented on the direct product space of the stochastic vector, so the network consists of units and states. The learning rule follows gradient decent formulation so as to minimize Kullback divergence between the output and the desired states. A preliminary recognition experiment on alphabetic characters was conducted, and SVM's internal representations were examined from weight patterns in the network. The experiment indicates that distributed or local representation is developed by the learning algorithm. A network system was constructed and applied to alphanumeric character recognition. Experimental results indicate that the SVM can perform as well as the BP model. >
9 citations
1 Jan 1989
TL;DR: The authors examine internal representations of a stochastic vector machine (SVM) and compare its learning ability with that of the backpropagation (BP) model through the experiments of handwritten Chinese character recognition, showing that local or distributed representation of input patterns is developed, depending on the structure of SVM.
Abstract: Summary form only given. The authors examine internal representations of a stochastic vector machine (SVM) and compare its learning ability with that of the backpropagation (BP) model through the experiments of handwritten Chinese character recognition. The experimental results show two facts. One is that local or distributed representation of input patterns is developed, depending on the structure of SVM. The other is that the recognition rate of the SVM is almost equal to that of the BP model, so its learning ability can be considered to be comparable to that of BP model. >
1 citations