|
1. |
Wafer‐scale integration implementation of mesh‐connected multiprocessor systems |
|
Systems and Computers in Japan,
Volume 26,
Issue 1,
1995,
Page 1-10
Issei Numata,
Susumu Horiguchi,
Preview
|
PDF (744KB)
|
|
摘要:
AbstractWith recent advances in VLSI technology, Wafer‐Scale Integration implementation is becoming realistic, wherein massively parallel multiprocessors are implemented on a single wafer. As a result of this technique, the connections among VLSI become unnecessary and a high‐speed, high‐density system will be realized. This paper considers the mesh‐connected network, which is one of the configurations of the massively parallel multiprocessor system, and discusses the reconfiguration techniques in realizing the network on the wafer. The reconfiguration technique is necessary for composing a system on a wafer, and several techniques have been proposed to date. Most of these, however, are based on global information or can be applied only to special cases. Consequently, they are not suitable for the self‐reconfiguration scheme of a massively parallel system. From such a viewpoint, this paper proposes and evaluates an autonomic self‐reconfiguration scheme using only local information, wherein the nondeterministic algorithm is added to the conventional method. It is seen as a result that the proposed method can realize a higher yield than the conventional method, as the array size becomes larger and the portion of the redundant PE b
ISSN:0882-1666
DOI:10.1002/scj.4690260101
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1995
数据来源: WILEY
|
2. |
A method for generating program specification from source program—analysis by transforming program structure and argument manipulation |
|
Systems and Computers in Japan,
Volume 26,
Issue 1,
1995,
Page 11-25
Takahiro Nagai,
Takeshi Imanaka,
Jun'Ichi Toyoda,
Tsukasa Hirashima,
Kuniaki Uehara,
Yuichi Nagasawa,
Preview
|
PDF (1214KB)
|
|
摘要:
AbstractThis paper proposes a method of transformation analysis for the automatic generation of program specification from the source program. The system APSG/I (Automatic Program Specification Generator I), on which the proposed method is implemented, is discussed. The transformation analysis notes the similarity to the list‐processing program of Prolog. By utilizing the framework of analogical reasoning, the specification for the given program is generated based on the typical list‐processing programs and its specification.When the program is inputted, the following two aspects are examined and the difference is determined by comparison with the already defined typical program: 1 the structure specified by the existence of the command or the order of the execution; and 2 the combination of arguments specified by the passing of data, etc. Then based on the determined difference, the already defined program specification is transformed, and the specification for the inputted program is generated. The specification generated by the proposed transformation analysis has the following features: 1 the result is the natural language sentence, since the program specification and the transformation rule for the specification are defined by the natural language; and 2 the format is unified and sufficient information to the details is contai
ISSN:0882-1666
DOI:10.1002/scj.4690260102
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1995
数据来源: WILEY
|
3. |
High‐speed match algorithm of a production system |
|
Systems and Computers in Japan,
Volume 26,
Issue 1,
1995,
Page 26-39
Haruhiko Kimura,
Shin‐Ya Kobayashi,
Tsuyoshi Takebe,
Kazuyuki Sumiyoshi,
Preview
|
PDF (1041KB)
|
|
摘要:
AbstractThe production system is suited to representing the miscellaneous pieces of human knowledge, and is used most widely in constructing the expert system. A problem is that a tremendous amount of time is required for the condition match, and an efficient condition match algorithm is desired.This paper considers the expert system which has been designed for the inference based on the semantic network, and aims at the realization of the high‐speed condition match in the expert system rewritten by OPS5. The working memory elements which are the objects of investigation are mostly the knowledge of two‐term relations, and the number of attributes usually is 2.This paper also proposes a data structure for the working memory, which can derive directly all working memory elements satisfying the condition, from the condition elements with 2 or less attributes. Using the proposed data structure, an algorithm is shown that can realize a high‐speed condition match. It is shown for the considered expert system that the proposed algorithm can realize a higher speed than the conventional condition match algo
ISSN:0882-1666
DOI:10.1002/scj.4690260103
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1995
数据来源: WILEY
|
4. |
A method of anaphora resolution based on the concept of observation |
|
Systems and Computers in Japan,
Volume 26,
Issue 1,
1995,
Page 40-49
Tsuyoshi Yamamura,
Noboru Ohnishi,
Noboru Sugie,
Preview
|
PDF (876KB)
|
|
摘要:
AbstractThe meaning of a sentence depends on the context in which it is used with other sentences. This results from the fact that several sentences are usually integrated to form a context. To date, the study of context structure has been concerned mostly with the semantic relation between the verbs (predicates) of the two sentences. It should be noted, however, that even if the verbs are semantically related, the relation may be felt to be unnatural and cannot be considered to form a context when the relation called discourse factor, such as the relationship between viewpoint, is not established. This study considers the context construction problem from the viewpoint of the natural relation between sentences. For this purpose, the concept of observation is introduced, this is an extension of viewpoint. Using that concept, it is shown how the interpretation of the sentence is helped by the preceding sentence. Then, the application to the resolution of anaphora is discussed. In other words, the anaphora is resolved by establishing the natural relation between sentences. It is shown that the anaphora problem which has not been resolved by the past method can now be resolved. The effectiveness of the anaphora resolution considered in this study is verified by examining actual examples.
ISSN:0882-1666
DOI:10.1002/scj.4690260104
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1995
数据来源: WILEY
|
5. |
Direct match algorithm for production systems |
|
Systems and Computers in Japan,
Volume 26,
Issue 1,
1995,
Page 50-61
Haruhiko Kimura,
Shin‐Ya Kobayashi,
Tsuyoshi Takebe,
Kazuyuki Sumiyoshi,
Preview
|
PDF (935KB)
|
|
摘要:
AbstractA production system is a typical tool for constructing an expert system. The Rete algorithm has been used as the condition match in the production system. The Rete algorithm features in that the condition match can executed efficiently by retaining the intermediate result of matching using the data‐flow graph (called Rete network). Efficiency is deteriorated, however, when most of the intermediate results are updated by a single recognition‐action cycle. Moreover, execution time for the condition match increases rapidly with an increase in the working memory elements. From such a viewpoint, this paper proposes a condition match algorithm wherein the intermediate result of matching is not retained, and the number of working memory tuples to be examined is reduced by connecting the common variables among the condition element. The proposed algorithm directly tries the condition match without using the Rete algorithm. The execution time does not increase so rapidly as in the Rete algorithm, but increases only slowly with the increase of the working memory elements. Consequently, the efficiency of the Rete algorithm is better when there are fewer working memory elements, and the efficiency of the proposed method is better when both the number of working memory elements and the computation cost for the condition match are increased. In other words, the proposed algorithm can handle high‐cost
ISSN:0882-1666
DOI:10.1002/scj.4690260105
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1995
数据来源: WILEY
|
6. |
Restoration of shift variant blurred image estimating the parameter distribution of point‐spread function |
|
Systems and Computers in Japan,
Volume 26,
Issue 1,
1995,
Page 62-72
Shoichi Hashimoto,
Hideo Saito,
Preview
|
PDF (1526KB)
|
|
摘要:
AbstractThis paper describes a method of restoring a position‐dependent (shift variant) blurred image by estimating the size of the blurred part and by adjusting the restoration proportionally to it. In this method, it is assumed that all blurred images are due to a defocussing of a lens system or accidental camera motion. If the point‐spread function (PSF) of the blurred point is known, the estimation of the PSF can be replaced by the estimation of its parameter distribution (consisting of the diameter of a defocussed circle and the width of the motion caused by the accidental camera motions). Then a blind deconvolution of a shift‐variant blurred image can be realized by locally applying a Wiener filter to the estimated distribution. The method was tested using computer simulated images and actual images taken by a CCD camera, both with blurred parts. The results confirm the effectiveness of the m
ISSN:0882-1666
DOI:10.1002/scj.4690260106
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1995
数据来源: WILEY
|
7. |
A generalized algorithm for Japanese morphological analysis and a comparative evaluation of some heuristics |
|
Systems and Computers in Japan,
Volume 26,
Issue 1,
1995,
Page 73-87
Toru Hisamitsu,
Yoshihiko Nitta,
Preview
|
PDF (995KB)
|
|
摘要:
AbstractIn ordinary written Japanese, words are not separated by spaces. Therefore morphological analysis involves segmenting and tagging sentences. Since each sentence has a huge number of possible tagged segmentations, various criteria have been proposed for making plausible decisions. However, there are still no unified frameworks that incorporate various heuristics, and there has been no comparative evaluation of commonly used heuristics. This paper presents a clear framework to describe various heuristics, and an N‐best algorithm for extracting optimal solutions. The time complexity of this algorithm isO(nNlog2(1 +N)), wherenis the sentence length. The advantage of the N‐best algorithm over the standard beam search algorithm is also discussed. This paper also presents a comparative evaluation of three major heuristics, and proposes a precise and portable rule‐based heuristic. Estimation was done using the aforementioned algorithm and six criteria. The newly proposed heuristic is based upon the Extended Least Bunsetsu (Phrase) Number m
ISSN:0882-1666
DOI:10.1002/scj.4690260107
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1995
数据来源: WILEY
|
8. |
Artifact cancellation in MRI due to phase encoding axis motion |
|
Systems and Computers in Japan,
Volume 26,
Issue 1,
1995,
Page 88-98
Li Tang,
Muneki Ohya,
Yoshinobu Sato,
Shinichi Tamura,
Hiroaki Naito,
Koushi Harada,
Takahiro Kozuka,
Preview
|
PDF (1373KB)
|
|
摘要:
AbstractThis paper addresses the problem of motion artifact cancellation in the standard 2D‐FFT MRI. Although motion is possible in any direction, in clinical MRI diagnosis, respiration is considered to be the main cause of motion. As a first order of approximation, motion due to respiration is assumed to be only in the phase‐encoding direction (Y‐axis) in this paper. In the previous approach, the region of a target object is assumed to be known and an iterative procedure is required. The problem of the algorithm is the convergence of the iterative (which may take a lot of time) and still has no guarantee of convergence. To avoid an iterative procedure, a new constraint is proposed here, with which the motion component and the true image component can be separated by a simple algebraic operation. After the Fourier transform of MRI signal along the read‐out direction (X‐direction), the phase of the spectrum alongY‐axis is expressed as an algebraic sum of the motion component and the true image component. On the other hand, when density distribution is symmetric along a Y‐direction line, such as a slice line on subcutaneous fat, the phase of the Fourier spectrum along the line is a linear function ofY‐position. Therefore, if the density is symmetric, the departure from the linear function of the phase is just the motion component. With this constraint the motion component can be estimated and the motion artifact in MRI can be canceled. The algorithm is shown to be effective using a phantom with simulated motions in various cases. When the density distribution along theY‐directional line on subcutaneous fat is not perfectly symmetric, the accuracy of estimated motion will be affected to some extent. However, this method is shown to be still effective in such general ca
ISSN:0882-1666
DOI:10.1002/scj.4690260108
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1995
数据来源: WILEY
|
9. |
A multimodule neural network model and the estimate of its nature as associative memory |
|
Systems and Computers in Japan,
Volume 26,
Issue 1,
1995,
Page 99-110
Seichi Ozawa,
Kazuyoshi Tsutsumi,
Preview
|
PDF (1056KB)
|
|
摘要:
AbstractA modular neural network model, “Cross‐Coupled Hopfield Nets with Many‐to‐Many Mapping Internetworks (CCHN‐MMMI),” in which the total consumed energy is defined as a linear sum of energy functions for subnetworks and their interactions, is proposed herein. This model can be applied to cases in which many‐to‐many mappings are necessary to mutually relate states of modules. The network dynamics are derived first, and then through computer simulations, the characteristics of the 2‐module CCHN‐MMMI are studied as associative memory. In one simulation, the effects of the introduction of explicitly modular architecture and a comparison of the results to those produced by conventional autocorrelation associative memory models.Another simulation shows that the model can store and recall pairs of character patterns successfully even if many‐to‐many mapping relations are necessary to relate the pairs to each other. A quantitative investigation as to how the recall dynamics deteriorate with an increasing number of randomly chosen fundamental memories is made in order to access the model as associative. The results indicate that internetworks between modules function effectively to prevent the entire network from recalling spurious memories. When multilayered networks with nonlinear hidden units are employed as internetworks in the model, the association performance is even more greatly improved as compared with conventional autocorrelation a
ISSN:0882-1666
DOI:10.1002/scj.4690260109
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1995
数据来源: WILEY
|
|