Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218512
R. Kamimura
A method of accelerating learning in recurrent neural networks is described. To activate and use connections, a complexity term defined by an equation was added to a standard quadratic error function. In experiments, a method was used in which a derivative of the complexity term was normally effective for positive connections, while negative connections were pushed toward smaller values. Thus, some connections were expected to be activated and were large enough to speed up learning. It was confirmed that the complexity term was effective in increasing the variance of the connections, especially the hidden connections. It was also confirmed that eventually some connections, especially some hidden connections, were activated and were large enough to be used in speeding up learning.<>
{"title":"Activation of connections to accelerate the learning in recurrent back-propagation","authors":"R. Kamimura","doi":"10.1109/CMPEUR.1992.218512","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218512","url":null,"abstract":"A method of accelerating learning in recurrent neural networks is described. To activate and use connections, a complexity term defined by an equation was added to a standard quadratic error function. In experiments, a method was used in which a derivative of the complexity term was normally effective for positive connections, while negative connections were pushed toward smaller values. Thus, some connections were expected to be activated and were large enough to speed up learning. It was confirmed that the complexity term was effective in increasing the variance of the connections, especially the hidden connections. It was also confirmed that eventually some connections, especially some hidden connections, were activated and were large enough to be used in speeding up learning.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114809202","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218520
C. Alippi, R. Cucchiara
A classification of data by using the genetic algorithm computational paradigm is proposed. The best data partition is defined to be the one minimizing the sum of Pythagorean distances between each datum in a cluster and the relative center of class or center of mass. Background is given, and the relevant genetic algorithm description is provided. The model for the genetic application is presented. Simulation results confirm genetic algorithms to be powerful tools for the solution of optimization problems.<>
{"title":"Cluster partitioning in image analysis classification: a genetic algorithm approach","authors":"C. Alippi, R. Cucchiara","doi":"10.1109/CMPEUR.1992.218520","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218520","url":null,"abstract":"A classification of data by using the genetic algorithm computational paradigm is proposed. The best data partition is defined to be the one minimizing the sum of Pythagorean distances between each datum in a cluster and the relative center of class or center of mass. Background is given, and the relevant genetic algorithm description is provided. The model for the genetic application is presented. Simulation results confirm genetic algorithms to be powerful tools for the solution of optimization problems.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123219818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218505
F. Wagner
A software design method is presented, based on a virtual finite state machine (VFSM) concept. The concept defines a virtual environment that allows the finite state machine to be an entirely table driven software module. A hybrid finite state model is used to achieve a superior design clarity. The method separates the control part of the design problem from the data manipulation part. The specification of the control part is directly executable. The control part is not coded. It is expressed in a table that is executed by the virtual finite state machine executor. The concept of virtual environment has been utilized for the design of complex software control systems.<>
{"title":"VFSM executable specification","authors":"F. Wagner","doi":"10.1109/CMPEUR.1992.218505","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218505","url":null,"abstract":"A software design method is presented, based on a virtual finite state machine (VFSM) concept. The concept defines a virtual environment that allows the finite state machine to be an entirely table driven software module. A hybrid finite state model is used to achieve a superior design clarity. The method separates the control part of the design problem from the data manipulation part. The specification of the control part is directly executable. The control part is not coded. It is expressed in a table that is executed by the virtual finite state machine executor. The concept of virtual environment has been utilized for the design of complex software control systems.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123271187","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218438
E. Schoneburg
Diagnosis is an important part of the product quality assurance cycle. The construction of diagnosis systems with neural nets, which solves problems hitherto unsolved, creates effective real-time opportunities, and a new level of robustness is discussed. The application of neural nets is illustrated in three examples. The first example shows how the effect of medication can be diagnosed. The other two involve automotive applications. They were derived from industrial products created by Expert Informatik GmbH over the years 1989 to 1992.<>
{"title":"Diagnosis using neural nets","authors":"E. Schoneburg","doi":"10.1109/CMPEUR.1992.218438","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218438","url":null,"abstract":"Diagnosis is an important part of the product quality assurance cycle. The construction of diagnosis systems with neural nets, which solves problems hitherto unsolved, creates effective real-time opportunities, and a new level of robustness is discussed. The application of neural nets is illustrated in three examples. The first example shows how the effect of medication can be diagnosed. The other two involve automotive applications. They were derived from industrial products created by Expert Informatik GmbH over the years 1989 to 1992.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"88 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129109833","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218433
J.R. Gonzalez de Mendivil, J. R. Garitagoitia, A. Córdoba
The authors introduce a simple and basic distributed deadlock detection and resolution algorithm for the one-resource model of a distributed database system. The distribution database model upon which the algorithm is based is the model introduced by D.A. Menasce and R.R. Muntz (1979). The main contribution is the use of automata theory for modeling the formal specification of the algorithm. This mathematical model allows the development of an elegant and short proof of correctness. The algorithm reports all true deadlocks and generates a unique candidate for resolving the deadlock.<>
{"title":"A simple distributed deadlock detection algorithm","authors":"J.R. Gonzalez de Mendivil, J. R. Garitagoitia, A. Córdoba","doi":"10.1109/CMPEUR.1992.218433","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218433","url":null,"abstract":"The authors introduce a simple and basic distributed deadlock detection and resolution algorithm for the one-resource model of a distributed database system. The distribution database model upon which the algorithm is based is the model introduced by D.A. Menasce and R.R. Muntz (1979). The main contribution is the use of automata theory for modeling the formal specification of the algorithm. This mathematical model allows the development of an elegant and short proof of correctness. The algorithm reports all true deadlocks and generates a unique candidate for resolving the deadlock.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130113282","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218521
N. Kagan, R. Adams
A computational system is presented to assist decision makers in the task of planning electrical distribution power systems. Three basic models are devised: a single-objective optimization model, with technical and economic considerations; a multiobjective model, which allows the consideration of such aspects as reliability and environmental impact; and a fuzzy mathematical programming model, considering fuzzy objectives and constraints. Some other solution methods are briefly presented and compared. The computational system implemented is detailed and illustrated through a case study.<>
{"title":"A computational decision support system for electrical distribution systems planning","authors":"N. Kagan, R. Adams","doi":"10.1109/CMPEUR.1992.218521","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218521","url":null,"abstract":"A computational system is presented to assist decision makers in the task of planning electrical distribution power systems. Three basic models are devised: a single-objective optimization model, with technical and economic considerations; a multiobjective model, which allows the consideration of such aspects as reliability and environmental impact; and a fuzzy mathematical programming model, considering fuzzy objectives and constraints. Some other solution methods are briefly presented and compared. The computational system implemented is detailed and illustrated through a case study.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125249700","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218515
C. Jesshope
The author describes work directed towards the implementation of a general purpose scalable parallel computer for scientific and engineering applications. The strategy has been to develop and define a portable software platform which can be used as a very abstract compiler target language for data parallel programs. It is suggested that such an approach not only promotes portability of programs between different classes of parallel computers but also provides a means of gaining efficient implementation on distributed memory computers by avoiding unnecessary synchronization, and by contributing toward parallel slackness.<>
{"title":"The F-code abstract machine and its implementation","authors":"C. Jesshope","doi":"10.1109/CMPEUR.1992.218515","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218515","url":null,"abstract":"The author describes work directed towards the implementation of a general purpose scalable parallel computer for scientific and engineering applications. The strategy has been to develop and define a portable software platform which can be used as a very abstract compiler target language for data parallel programs. It is suggested that such an approach not only promotes portability of programs between different classes of parallel computers but also provides a means of gaining efficient implementation on distributed memory computers by avoiding unnecessary synchronization, and by contributing toward parallel slackness.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130257724","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218447
Mohammad S. Obaidat, D. Abu-Saymeh
The authors study, analyze, and compare the performance of pattern recognition methods with various neural network techniques for ultrasonic transducer characterization. The characterization algorithms are discussed. A multilayer backpropagation neural network is developed for characterizing the transducers. It provided a misclassification rate of 6%. Two other multilayer neural networks, sum-of-products and a newly devised neural network called hybrid sum-of-products, had misclassification rates of 10% and 7%, respectively. The best pattern recognition technique for this application was found to be the perceptron, which provided a misclassification rate of 23%. The worst pattern recognition technique was found to be the Bayes theorem method, which provided a misclassification rate of 54%. The competitive learning technique provided poor results as compared to the K-means for preclustering.<>
{"title":"On the performance of neural networks and pattern recognition paradigms for classifying ultrasonic transducers","authors":"Mohammad S. Obaidat, D. Abu-Saymeh","doi":"10.1109/CMPEUR.1992.218447","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218447","url":null,"abstract":"The authors study, analyze, and compare the performance of pattern recognition methods with various neural network techniques for ultrasonic transducer characterization. The characterization algorithms are discussed. A multilayer backpropagation neural network is developed for characterizing the transducers. It provided a misclassification rate of 6%. Two other multilayer neural networks, sum-of-products and a newly devised neural network called hybrid sum-of-products, had misclassification rates of 10% and 7%, respectively. The best pattern recognition technique for this application was found to be the perceptron, which provided a misclassification rate of 23%. The worst pattern recognition technique was found to be the Bayes theorem method, which provided a misclassification rate of 54%. The competitive learning technique provided poor results as compared to the K-means for preclustering.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121300547","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218508
M. Kara, P. Jesty, T. Gough
The authors present the design and implementation of a software framework using the blackboard model. This blackboard-based framework is an abstract interface used for the development of cooperating schedulers in a distributed system. The significance of this work is threefold: first, the interesting features of the blackboard model was applied to the area of software modeling; second, an abstract interface was built to distinguish blackboard modeling issues from scheduling issues; and third, a blackboard-based scheduler was implemented efficiently using lightweight processes.<>
{"title":"A blackboard-based framework for the development of cooperating schedulers","authors":"M. Kara, P. Jesty, T. Gough","doi":"10.1109/CMPEUR.1992.218508","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218508","url":null,"abstract":"The authors present the design and implementation of a software framework using the blackboard model. This blackboard-based framework is an abstract interface used for the development of cooperating schedulers in a distributed system. The significance of this work is threefold: first, the interesting features of the blackboard model was applied to the area of software modeling; second, an abstract interface was built to distinguish blackboard modeling issues from scheduling issues; and third, a blackboard-based scheduler was implemented efficiently using lightweight processes.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116474051","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1992-05-04DOI: 10.1109/CMPEUR.1992.218415
S. Seinecke
Presents a basic scheme for water-cooled multichip modules (MCMs). Multichip modules are a very effective approach to achieve high packaging density, excellent cooling and high electrical performance of mainframe processors. Two characteristic approaches are compared, namely an MCM with flip-chip mounting and an MCM with micropacks. The electrical characteristics of chip-to-board connections, microwiring substrates and planar module connectors have to be carefully designed. The ICs should preferably be BiCMOS AS ICs with up to 80000 gate functions with optional embedded SRAMs of up to 360000-bit capacity. The mechanical construction, wiring and delay rules, and test requirements for such chips are considered.<>
提出了水冷多芯片模块(mcm)的基本方案。多芯片模块是一种非常有效的方法，以实现高封装密度，优良的冷却和高电气性能的主机处理器。比较了两种典型的MCM方法，即采用倒装芯片的MCM和采用微封装的MCM。芯片到板连接、微布线基板和平面模块连接器的电气特性必须仔细设计。ic最好是具有多达80000门功能的BiCMOS AS ic，可选的嵌入式ram容量高达360,000位。考虑了这种芯片的机械结构、布线和延迟规则以及测试要求。
{"title":"High-performance MCMs for mainframe computers","authors":"S. Seinecke","doi":"10.1109/CMPEUR.1992.218415","DOIUrl":"https://doi.org/10.1109/CMPEUR.1992.218415","url":null,"abstract":"Presents a basic scheme for water-cooled multichip modules (MCMs). Multichip modules are a very effective approach to achieve high packaging density, excellent cooling and high electrical performance of mainframe processors. Two characteristic approaches are compared, namely an MCM with flip-chip mounting and an MCM with micropacks. The electrical characteristics of chip-to-board connections, microwiring substrates and planar module connectors have to be carefully designed. The ICs should preferably be BiCMOS AS ICs with up to 80000 gate functions with optional embedded SRAMs of up to 360000-bit capacity. The mechanical construction, wiring and delay rules, and test requirements for such chips are considered.<<ETX>>","PeriodicalId":390273,"journal":{"name":"CompEuro 1992 Proceedings Computer Systems and Software Engineering","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1992-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126532302","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：