A universal digital VLSI design for implementing artificial neural networks is proposed. The design is based on a unified iterative neural network model. An implementation based on a combination of custom-built and commercially available chips is presented. In addition, a software environment for the array processor is discussed.<>
{"title":"Special purpose array processor implementation of neural networks","authors":"S. Kung, J. Vlontzos","doi":"10.1109/TAI.1990.130322","DOIUrl":"https://doi.org/10.1109/TAI.1990.130322","url":null,"abstract":"A universal digital VLSI design for implementing artificial neural networks is proposed. The design is based on a unified iterative neural network model. An implementation based on a combination of custom-built and commercially available chips is presented. In addition, a software environment for the array processor is discussed.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116847054","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}
PROSE, a generic package for maintaining the consistency of object databases using constraint driven computations is discussed. Two original contributions of the package which arose from the meeting of objects and constraints are described. First, due to its reflective implementation, PROSE makes it possible keep consistency not only for the database entities but also for the constraint networks it builds. Second, it provides the user with temporal constraints that help supervise the evolution of the entities between arbitrary time events. A validation of this approach is shown by describing an instantiation of PROSE for an expert system shell.<>
{"title":"Constraint driven computations in evolving object databases","authors":"Pierre Berlandier","doi":"10.1109/TAI.1990.130328","DOIUrl":"https://doi.org/10.1109/TAI.1990.130328","url":null,"abstract":"PROSE, a generic package for maintaining the consistency of object databases using constraint driven computations is discussed. Two original contributions of the package which arose from the meeting of objects and constraints are described. First, due to its reflective implementation, PROSE makes it possible keep consistency not only for the database entities but also for the constraint networks it builds. Second, it provides the user with temporal constraints that help supervise the evolution of the entities between arbitrary time events. A validation of this approach is shown by describing an instantiation of PROSE for an expert system shell.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114343669","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}
An improved adaptable unification order algorithm is presented. It is used in register allocation for a Warren abstract machine (WAM) Prolog implementation. The algorithm is based on the division of a clause into chunks. The purpose is to generate better Prolog codes for programs so that they have fewer instructions and data movements in execution. It is realized by a good conflict handling algorithm and by an adaptable processing order method of compound arguments. The method has been proved to be efficient in practice.<>
{"title":"An improved register allocation algorithm for the WAM model","authors":"Hu Ziang, Xiang Rong Ci","doi":"10.1109/TAI.1990.130415","DOIUrl":"https://doi.org/10.1109/TAI.1990.130415","url":null,"abstract":"An improved adaptable unification order algorithm is presented. It is used in register allocation for a Warren abstract machine (WAM) Prolog implementation. The algorithm is based on the division of a clause into chunks. The purpose is to generate better Prolog codes for programs so that they have fewer instructions and data movements in execution. It is realized by a good conflict handling algorithm and by an adaptable processing order method of compound arguments. The method has been proved to be efficient in practice.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114761511","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}
Scaffolding is a novel method for improving conjunctive match efficiency that incorporates a TMS (truth maintenance system) into an RETE matcher, thereby exploiting match redundancy between separate runs. The resulting inactivations are reactivations of the data dependencies replace expensive RETE testing with simpler, less costly NETL-like propagations. Experimental results obtained by profiling with a non-toy expert system substantiate scaffolding's between-trial transfer capability. The experiments demonstrate that there can be significant transfer of match knowledge between runs on different, homologous problems. The construction and retrieval of these matches are efficiently built into the algorithm. Scaffolding's utility and relation to other artificial-intelligence methods are discussed.<>
{"title":"Scaffolding the RETE network","authors":"M. Perlin","doi":"10.1109/TAI.1990.130367","DOIUrl":"https://doi.org/10.1109/TAI.1990.130367","url":null,"abstract":"Scaffolding is a novel method for improving conjunctive match efficiency that incorporates a TMS (truth maintenance system) into an RETE matcher, thereby exploiting match redundancy between separate runs. The resulting inactivations are reactivations of the data dependencies replace expensive RETE testing with simpler, less costly NETL-like propagations. Experimental results obtained by profiling with a non-toy expert system substantiate scaffolding's between-trial transfer capability. The experiments demonstrate that there can be significant transfer of match knowledge between runs on different, homologous problems. The construction and retrieval of these matches are efficiently built into the algorithm. Scaffolding's utility and relation to other artificial-intelligence methods are discussed.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125320958","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}
A suite of software, ASLANN, which facilitates the generation and simulation of composite neural systems formed from assemblies of neural and nonneural modules is described. Novel models for leaf modules and dedicated specialist display routines can be readily incorporated into the simulation of a neural system by linking with the simulator appropriate routines written in C. The tool has been used to produce a composite artificial neural system which performs the function of an integrated circuit channel router.<>
{"title":"ASLANN: a tool for the design of composite neural systems","authors":"A. Green, P. Noakes","doi":"10.1109/TAI.1990.130321","DOIUrl":"https://doi.org/10.1109/TAI.1990.130321","url":null,"abstract":"A suite of software, ASLANN, which facilitates the generation and simulation of composite neural systems formed from assemblies of neural and nonneural modules is described. Novel models for leaf modules and dedicated specialist display routines can be readily incorporated into the simulation of a neural system by linking with the simulator appropriate routines written in C. The tool has been used to produce a composite artificial neural system which performs the function of an integrated circuit channel router.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132234480","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}
Regularity is formalized from the general observation that hierarchical relationships between concepts reflect relationships between the concepts' properties, and vice-versa. A procedure called expansion is utilized to compute property values when those are not known, based on the patterns of regularity. Inheriting property values in taxonomies is a special case of expansion. Expansion may fail in tangled hierarchies in the same way that multiple inheritance may lead to conflicting inferences in taxonomies. The definition of regularity and the expansion procedure have been extended to fuzzy sets. The fuzzy sets framework takes into account the normative nature of property values and handles cases where expansion fails. A procedure parametrically enlarges sets of property values so that expansion does not fail. A theorem describes the effect of the choice of the enlargement parameter on the reliability of expansion. Applications of regularity and expansion to the building and maintenance of hierarchies are discussed.<>
{"title":"Regularity: generalizing inheritance to arbitrary hierarchies","authors":"H. Mili, R. Rada","doi":"10.1109/TAI.1990.130410","DOIUrl":"https://doi.org/10.1109/TAI.1990.130410","url":null,"abstract":"Regularity is formalized from the general observation that hierarchical relationships between concepts reflect relationships between the concepts' properties, and vice-versa. A procedure called expansion is utilized to compute property values when those are not known, based on the patterns of regularity. Inheriting property values in taxonomies is a special case of expansion. Expansion may fail in tangled hierarchies in the same way that multiple inheritance may lead to conflicting inferences in taxonomies. The definition of regularity and the expansion procedure have been extended to fuzzy sets. The fuzzy sets framework takes into account the normative nature of property values and handles cases where expansion fails. A procedure parametrically enlarges sets of property values so that expansion does not fail. A theorem describes the effect of the choice of the enlargement parameter on the reliability of expansion. Applications of regularity and expansion to the building and maintenance of hierarchies are discussed.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114571208","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}
It is shown that, in a hierarchically structured hypotheses space, any belief function whose focal elements are nodes in the hierarchy is a separable support function. An algorithm is proposed that decomposes such a separable support function into simple support functions. It is shown that the computational complexity of this decomposition algorithm is O(N/sup 2/). Applications of the decomposition of separable support functions to the data fusion problem and the reasoning about control problem are discussed.<>
{"title":"Decomposition of belief function in hierarchical hypotheses space","authors":"H. Hau","doi":"10.1109/TAI.1990.130427","DOIUrl":"https://doi.org/10.1109/TAI.1990.130427","url":null,"abstract":"It is shown that, in a hierarchically structured hypotheses space, any belief function whose focal elements are nodes in the hierarchy is a separable support function. An algorithm is proposed that decomposes such a separable support function into simple support functions. It is shown that the computational complexity of this decomposition algorithm is O(N/sup 2/). Applications of the decomposition of separable support functions to the data fusion problem and the reasoning about control problem are discussed.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128221966","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}
Consideration is given to EDS/OWL, an object-oriented environment which provides a uniform framework for integrating programming paradigms. The authors describe how EDS/OWL integrates access-oriented, rule-based and logic programming in an object-oriented environment. The integration extends the underlying language yet continues to use the existing constructs as well as the parser, interpreter and tools. Moreover, the object-oriented framework supports further novel extensions frequently required by specific artificial-intelligence (AI) applications.<>
{"title":"An object-oriented environment for multiple AI paradigms","authors":"M. Ibrahim, S.W. Woyak","doi":"10.1109/TAI.1990.130313","DOIUrl":"https://doi.org/10.1109/TAI.1990.130313","url":null,"abstract":"Consideration is given to EDS/OWL, an object-oriented environment which provides a uniform framework for integrating programming paradigms. The authors describe how EDS/OWL integrates access-oriented, rule-based and logic programming in an object-oriented environment. The integration extends the underlying language yet continues to use the existing constructs as well as the parser, interpreter and tools. Moreover, the object-oriented framework supports further novel extensions frequently required by specific artificial-intelligence (AI) applications.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132385388","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}
With the long-range goal of developing a practical methodology for complete verification and validation of expert systems, the authors have investigated methods for designing knowledge bases so that they can be easily verified. These methods are based on extensions of traditional software verification technology and on emerging artificial intelligence programming technology. After discussing the background of this work, the authors describe the methods and their benefits to verification and discuss ongoing and future research activities.<>
{"title":"Designing verifiable expert systems","authors":"J. Yen, Jonathan Lee, David Hamilton","doi":"10.1109/TAI.1990.130453","DOIUrl":"https://doi.org/10.1109/TAI.1990.130453","url":null,"abstract":"With the long-range goal of developing a practical methodology for complete verification and validation of expert systems, the authors have investigated methods for designing knowledge bases so that they can be easily verified. These methods are based on extensions of traditional software verification technology and on emerging artificial intelligence programming technology. After discussing the background of this work, the authors describe the methods and their benefits to verification and discuss ongoing and future research activities.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133177567","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}
An object-oriented knowledge base management system (OKBMS) which has the ability to flexibly define and extend the object types has been developed. It provides a development environment for intelligent applications. The outstanding characteristics of object-oriented style such as abstract data type, information hiding, inheritance and message passing are fully implemented in the OKBMS. The OKBMS inference engine not only handles the certain reasoning but also can deal with uncertain information. In OKBMS, a semi-interpreted approach is adopted to integrate Prolog with the ORACLE database. The key issues involved in OKBMS, such as knowledge structure, inference algorithms, partial evaluation and comprehensive user interfaces are discussed.<>
{"title":"OKBMS: An object-oriented knowledge base management system","authors":"Zhongzhi Shi, Jianming Wu, Hui Sun, Jinxi Xu","doi":"10.1109/TAI.1990.130327","DOIUrl":"https://doi.org/10.1109/TAI.1990.130327","url":null,"abstract":"An object-oriented knowledge base management system (OKBMS) which has the ability to flexibly define and extend the object types has been developed. It provides a development environment for intelligent applications. The outstanding characteristics of object-oriented style such as abstract data type, information hiding, inheritance and message passing are fully implemented in the OKBMS. The OKBMS inference engine not only handles the certain reasoning but also can deal with uncertain information. In OKBMS, a semi-interpreted approach is adopted to integrate Prolog with the ORACLE database. The key issues involved in OKBMS, such as knowledge structure, inference algorithms, partial evaluation and comprehensive user interfaces are discussed.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1990-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125757845","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：