TIMEX, a tool that extends Prolog with features for interval-based representation, is described. Because several ideas for interval-based representation exist, the tool provides the possibility of switching between different representation and propagation techniques. This is done by representing different interval graphs. Thus, one application may consist of several graphs with different attributes for propagation and one interval may exist in different graphs. The application domain is technical expert systems. An interval-based representation was used for a scheduling expert system in a steelmaking plant.<>
{"title":"TIMEX-a tool for interval-based representation in technical applications","authors":"J. Dorn","doi":"10.1109/TAI.1990.130388","DOIUrl":"https://doi.org/10.1109/TAI.1990.130388","url":null,"abstract":"TIMEX, a tool that extends Prolog with features for interval-based representation, is described. Because several ideas for interval-based representation exist, the tool provides the possibility of switching between different representation and propagation techniques. This is done by representing different interval graphs. Thus, one application may consist of several graphs with different attributes for propagation and one interval may exist in different graphs. The application domain is technical expert systems. An interval-based representation was used for a scheduling expert system in a steelmaking plant.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"120 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":"114562475","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}
The authors have designed an artificial neural system (ANS) consisting of coupled back-error propagation (BEP) networks that perform feature extraction, clustering. and categorization of tactile surface impressions. The network and its characteristics are reviewed, with particular focus on its performance in the presence of noisy input patterns. Simulation results indicate that, regarding geometry-size- and activation-constrained grey-scale patterns, the BEP classifier is sensitive to both additive low-amplitude spike noise and additive white Gaussian noise. Most of the misclassifications occur among patterns that differ only by small variations in force gradients. The network's performance gradually improves when noisy patterns are included in the training set, but indications are that large training sets or alternative error functions in BEP will be required to achieve robust performance in the tactile domain.<>
{"title":"A study of back-error propagation networks in the domain of noisy tactile impressions","authors":"M. Thint, Paul P. Wang","doi":"10.1109/TAI.1990.130422","DOIUrl":"https://doi.org/10.1109/TAI.1990.130422","url":null,"abstract":"The authors have designed an artificial neural system (ANS) consisting of coupled back-error propagation (BEP) networks that perform feature extraction, clustering. and categorization of tactile surface impressions. The network and its characteristics are reviewed, with particular focus on its performance in the presence of noisy input patterns. Simulation results indicate that, regarding geometry-size- and activation-constrained grey-scale patterns, the BEP classifier is sensitive to both additive low-amplitude spike noise and additive white Gaussian noise. Most of the misclassifications occur among patterns that differ only by small variations in force gradients. The network's performance gradually improves when noisy patterns are included in the training set, but indications are that large training sets or alternative error functions in BEP will be required to achieve robust performance in the tactile domain.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"68 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":"121554694","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}
The problem of finding a collision-free path for a shape-changeable mobile robot in a 3D environment is studied. In order to solve the problem, a 3D path planner is used. This path planner represents the spaces which are occupied by the robot under different gestures as a set of cuboids of different sizes and uses two levels of planning, i.e. global path planning and local path planning to plan the path for the robot. The global path planner generates a most plausible global path for the robot. Subsequently, the local path planner first represents the free space along the plausible global path by a set of hexahedrons and then generates a connection graph whose vertices are the states of the robot and whose edges are operations of the robot by an expert system. Finally, a collision-free path is searched for in this connection graph.<>
{"title":"Collision-free path planning for a shape-changeable mobile robot in a 3-dimensional environment","authors":"Q. Xue, P. Sheu","doi":"10.1109/TAI.1990.130302","DOIUrl":"https://doi.org/10.1109/TAI.1990.130302","url":null,"abstract":"The problem of finding a collision-free path for a shape-changeable mobile robot in a 3D environment is studied. In order to solve the problem, a 3D path planner is used. This path planner represents the spaces which are occupied by the robot under different gestures as a set of cuboids of different sizes and uses two levels of planning, i.e. global path planning and local path planning to plan the path for the robot. The global path planner generates a most plausible global path for the robot. Subsequently, the local path planner first represents the free space along the plausible global path by a set of hexahedrons and then generates a connection graph whose vertices are the states of the robot and whose edges are operations of the robot by an expert system. Finally, a collision-free path is searched for in this connection graph.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"10 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":"132368318","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}
The authors propose a three-layered constraint satisfaction neural network to perform Huffman-Clowes scene labeling. Given a line drawing the network establishes a consistent labeling for all the edges or detects that it is physically unrealizable. Experimental results show that this approach exploits the parallel architecture inherent in the network and is faster than the conventional algorithmic method.<>
{"title":"Constraint propagation neural networks for Huffman-Clowes scene labeling","authors":"E. Tsao, Wei-Chung Lin","doi":"10.1109/TAI.1990.130345","DOIUrl":"https://doi.org/10.1109/TAI.1990.130345","url":null,"abstract":"The authors propose a three-layered constraint satisfaction neural network to perform Huffman-Clowes scene labeling. Given a line drawing the network establishes a consistent labeling for all the edges or detects that it is physically unrealizable. Experimental results show that this approach exploits the parallel architecture inherent in the network and is faster than the conventional algorithmic method.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"35 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":"132834723","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 hybrid knowledge engineering development environment called KEDE is briefly introduced as a powerful tool for large artificial-intelligence problems. It provides five kinds of knowledge representation: extended frames, semantic nets, procedural knowledge, object-oriented technique and predicate logic. Correspondingly, it supports procedure-oriented, data-oriented, object-oriented and logic-oriented programming. KEDE gains a very powerful inheritance mechanism using frames. It further provides an automatic retrieval technique for processing implicit knowledge, a demon mechanism for firing predicates or functions, a message-sending mechanism for activating methods, and two inference engines for backward, forward and mixed reasoning.<>
{"title":"An introduction to KEDE-A hybrid knowledge engineering development environment","authors":"Zijian Zheng, Wei Li","doi":"10.1109/TAI.1990.130311","DOIUrl":"https://doi.org/10.1109/TAI.1990.130311","url":null,"abstract":"A hybrid knowledge engineering development environment called KEDE is briefly introduced as a powerful tool for large artificial-intelligence problems. It provides five kinds of knowledge representation: extended frames, semantic nets, procedural knowledge, object-oriented technique and predicate logic. Correspondingly, it supports procedure-oriented, data-oriented, object-oriented and logic-oriented programming. KEDE gains a very powerful inheritance mechanism using frames. It further provides an automatic retrieval technique for processing implicit knowledge, a demon mechanism for firing predicates or functions, a message-sending mechanism for activating methods, and two inference engines for backward, forward and mixed reasoning.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"150 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":"123220894","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}
The authors provide a survey of the artificial intelligence (AI) R&D projects undertaken by the Institute of Systems Science (ISS) of the National University of Singapore. They also describe ISS's AI R&D strategies as well as the AI tools developed for various industrial projects such as the machine translation project with IBM, the connectionist troubleshooting expert system for an inertial navigation system with Singapore airlines, and the fuzzy pattern recognition system for identifying container numbers with the port authority of Singapore.<>
{"title":"Artificial intelligence R&D: the ISS strategy","authors":"J. Motiwalla, T. Heng","doi":"10.1109/TAI.1990.130414","DOIUrl":"https://doi.org/10.1109/TAI.1990.130414","url":null,"abstract":"The authors provide a survey of the artificial intelligence (AI) R&D projects undertaken by the Institute of Systems Science (ISS) of the National University of Singapore. They also describe ISS's AI R&D strategies as well as the AI tools developed for various industrial projects such as the machine translation project with IBM, the connectionist troubleshooting expert system for an inertial navigation system with Singapore airlines, and the fuzzy pattern recognition system for identifying container numbers with the port authority of Singapore.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"41 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":"122481502","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 genetic-algorithm-based methodology for learning a set of feature detectors for texture discrimination is presented. This methodology is incorporated into a vision system that learns to classify noisy examples of different texture classes by evolving populations of simple and texture-specific local spatial feature detectors. The results of preliminary work confirm the utility of the genetic algorithm in solving problems within the specified domain.<>
{"title":"Generation of feature detectors for texture discrimination by genetic search","authors":"J. Bala, K. A. Jong","doi":"10.1109/TAI.1990.130443","DOIUrl":"https://doi.org/10.1109/TAI.1990.130443","url":null,"abstract":"A genetic-algorithm-based methodology for learning a set of feature detectors for texture discrimination is presented. This methodology is incorporated into a vision system that learns to classify noisy examples of different texture classes by evolving populations of simple and texture-specific local spatial feature detectors. The results of preliminary work confirm the utility of the genetic algorithm in solving problems within the specified domain.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"35 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":"114169659","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}
The authors present a novel algorithm, speculative branch and bound, for performing heuristic search in restricted memory. Speculative branch-and-bound combines elements of best-first and depth-first searches. The authors describe the search and a variant, SIDA* that is analogous to A*. They show how they are related to such searches as IDA* and MA* show how to add pruning to SIDA* and give performance results. SIDA* is superior to IDA* and competitive with MA* in the number of nodes it expands, and its overhead lies between theirs.<>
{"title":"Speculative search: An efficient search algorithm for limited memory","authors":"M. Gooley, B. Wah","doi":"10.1109/TAI.1990.130333","DOIUrl":"https://doi.org/10.1109/TAI.1990.130333","url":null,"abstract":"The authors present a novel algorithm, speculative branch and bound, for performing heuristic search in restricted memory. Speculative branch-and-bound combines elements of best-first and depth-first searches. The authors describe the search and a variant, SIDA* that is analogous to A*. They show how they are related to such searches as IDA* and MA* show how to add pruning to SIDA* and give performance results. SIDA* is superior to IDA* and competitive with MA* in the number of nodes it expands, and its overhead lies between theirs.<<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":"121159900","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 tool system which supports the development of expert systems for use in the field of industrial automation is presented. The knowledge representation language provided by the tool system supports different aspects of knowledge which are important in this domain: the representation of declarative and procedural knowledge; the representation of temporal knowledge; and the representation of knowledge about real-time behavior. As a further part of the tool system, an expert system compiler with the ability to transform expert systems into a conventional procedural program (realized in C) is described. This method allows a simple integration of knowledge-based techniques into conventional software systems.<>
{"title":"A tool system for knowledge-based on-line diagnosis in industrial automation","authors":"T. Beck","doi":"10.1109/TAI.1990.130374","DOIUrl":"https://doi.org/10.1109/TAI.1990.130374","url":null,"abstract":"A tool system which supports the development of expert systems for use in the field of industrial automation is presented. The knowledge representation language provided by the tool system supports different aspects of knowledge which are important in this domain: the representation of declarative and procedural knowledge; the representation of temporal knowledge; and the representation of knowledge about real-time behavior. As a further part of the tool system, an expert system compiler with the ability to transform expert systems into a conventional procedural program (realized in C) is described. This method allows a simple integration of knowledge-based techniques into conventional software systems.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"418 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":"116683869","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}
Efficient ways to prioritize and gather evidence within belief networks are discussed. The authors also suggest ways in which one can structure a large problem (a ship classification problem in the present case) into a series of small ones. This both re-defines much of the control strategy into the system structure and also localizes run-time control issues into much smaller networks. The overall control strategy thus includes the combination of both of these methods. By combining them correctly one can reduce the amount of dynamic computation required during run-time, and thus improve the responsiveness of the system. When dealing with the ship classification problem, the techniques described appear to work well.<>
{"title":"A real time control strategy for Bayesian belief networks with application to ship classification problem solving","authors":"S. Musman, LiWu Chang, L. Booker","doi":"10.1109/TAI.1990.130430","DOIUrl":"https://doi.org/10.1109/TAI.1990.130430","url":null,"abstract":"Efficient ways to prioritize and gather evidence within belief networks are discussed. The authors also suggest ways in which one can structure a large problem (a ship classification problem in the present case) into a series of small ones. This both re-defines much of the control strategy into the system structure and also localizes run-time control issues into much smaller networks. The overall control strategy thus includes the combination of both of these methods. By combining them correctly one can reduce the amount of dynamic computation required during run-time, and thus improve the responsiveness of the system. When dealing with the ship classification problem, the techniques described appear to work well.<<ETX>>","PeriodicalId":366276,"journal":{"name":"[1990] Proceedings of the 2nd International IEEE Conference on Tools for Artificial Intelligence","volume":"23 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":"116712990","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：