A lot of research papers focus on the challenging problem of the combination of genetic algorithms and artificial neural networks. Developmental and molecular biology may be a source of inspiration for designing powerful artificial neurogenesis systems allowing the generation of complex modular structures. This paper describes in detail such a neurogenesis model associated with an evolutionary process and its application to the control of a mobile robot. Early results demonstrate the surprising efficiency of this methodology and give hints to continue the research towards the generation of more complex adaptive neural networks.
{"title":"Artificial neurogenesis: an application to autonomous robotics","authors":"O. Michel, P. Collard","doi":"10.1109/TAI.1996.560453","DOIUrl":"https://doi.org/10.1109/TAI.1996.560453","url":null,"abstract":"A lot of research papers focus on the challenging problem of the combination of genetic algorithms and artificial neural networks. Developmental and molecular biology may be a source of inspiration for designing powerful artificial neurogenesis systems allowing the generation of complex modular structures. This paper describes in detail such a neurogenesis model associated with an evolutionary process and its application to the control of a mobile robot. Early results demonstrate the surprising efficiency of this methodology and give hints to continue the research towards the generation of more complex adaptive neural networks.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"123 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121249617","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}
Ambiguity is one of the main sources of complexity in natural language processing. We propose an original solution relying on the use of named disjunctions which are set of ordered formulae linked by a name. This representation is implemented within a constraint logic programming paradigm (allowing the use of underspecified structures) by means of lazy evaluation techniques. This approach avoids the expansion of the disjunctions into a normal form and allows one, in some cases, to compute a partial syntactic structure without disambiguating.
{"title":"Named disjunctions and lazy evaluation for syntactic ambiguities","authors":"P. Blache","doi":"10.1109/TAI.1996.560785","DOIUrl":"https://doi.org/10.1109/TAI.1996.560785","url":null,"abstract":"Ambiguity is one of the main sources of complexity in natural language processing. We propose an original solution relying on the use of named disjunctions which are set of ordered formulae linked by a name. This representation is implemented within a constraint logic programming paradigm (allowing the use of underspecified structures) by means of lazy evaluation techniques. This approach avoids the expansion of the disjunctions into a normal form and allows one, in some cases, to compute a partial syntactic structure without disambiguating.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116153878","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}
Intelligent hybrid systems are playing an increasing role in the development of artificial intelligence. In this study, we applied simulated annealing to adjust the weights of a multilayer neural network (MNN). Two versions of simulated annealing were tested: conventional simulated annealing (CSA) and fast simulated annealing (FSA). The applied hybrid system was used as a classifier in order to discriminate between 3 seed species (1 cultivated seed species which is perennial rye grass, and 2 adventitious seed species which are rumex and wild oat). From a set of colour digital images, 73 morphometrical and textural features were extracted to characterise each individual seed. Stepwise discriminant analysis made it possible to select the first 3 relevant features. The performances of classification were highly dependent on the scaling parameters of simulated annealing. For example, when the number of iterations of simulated annealing was 5, and the number of temperatures was 40, the combination between CSA and MNN correctly classified 98.18 and 97.77 percent of the training and the test sets, whereas FSA and MNN identified 99.18 and 99.68 percent of the same data sets. Globally, FSA outperformed CSA both in reliability and computational resources. A hybrid system combined with a colour image analysis showed promise for the design of an automatic seed identification device.
{"title":"Application of a hybrid neural network for the discrimination of seeds by artificial vision","authors":"Y. Chtioui, D. Bertrand, M. Devaux, D. Barba","doi":"10.1109/TAI.1996.560797","DOIUrl":"https://doi.org/10.1109/TAI.1996.560797","url":null,"abstract":"Intelligent hybrid systems are playing an increasing role in the development of artificial intelligence. In this study, we applied simulated annealing to adjust the weights of a multilayer neural network (MNN). Two versions of simulated annealing were tested: conventional simulated annealing (CSA) and fast simulated annealing (FSA). The applied hybrid system was used as a classifier in order to discriminate between 3 seed species (1 cultivated seed species which is perennial rye grass, and 2 adventitious seed species which are rumex and wild oat). From a set of colour digital images, 73 morphometrical and textural features were extracted to characterise each individual seed. Stepwise discriminant analysis made it possible to select the first 3 relevant features. The performances of classification were highly dependent on the scaling parameters of simulated annealing. For example, when the number of iterations of simulated annealing was 5, and the number of temperatures was 40, the combination between CSA and MNN correctly classified 98.18 and 97.77 percent of the training and the test sets, whereas FSA and MNN identified 99.18 and 99.68 percent of the same data sets. Globally, FSA outperformed CSA both in reliability and computational resources. A hybrid system combined with a colour image analysis showed promise for the design of an automatic seed identification device.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114904892","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}
Recently, efficient algorithms have been proposed to achieve arc- and path-consistency in constraint networks. The best path-consistency algorithm proposed is PE-{5|6} which is a natural generalization of AC-6 to path-consistency independently proposed by M. Singh (1995) for PC-5 and A. Chmeiss and P. Jegou (1995) for PC-6. Unfortunately, we have remarked that PC-{5|6}, though it is widely better than PC-4 (Chmeiss and P. Jegou, 1996) was not very efficient in practice, especially for those classes of problems that require an important space to be run. So, we propose a new path-consistency algorithm called PC-8, the space complexity of which is O(n/sup 2/d) but its time complexity is O(n/sup 3/d/sup 4/), i.e. worse than that of PC-{5|6}. However, the simplicity of PC-8 as well as the data structures used for its implementation offer a higher performance than PC-{5|6}. The principle of PC-8 is also used to propose a new algorithm to achieve arc-consistency called AC-8.
{"title":"Two new constraint propagation algorithms requiring small space complexity","authors":"A. Chmeiss, Philippe Jégou","doi":"10.1109/TAI.1996.560465","DOIUrl":"https://doi.org/10.1109/TAI.1996.560465","url":null,"abstract":"Recently, efficient algorithms have been proposed to achieve arc- and path-consistency in constraint networks. The best path-consistency algorithm proposed is PE-{5|6} which is a natural generalization of AC-6 to path-consistency independently proposed by M. Singh (1995) for PC-5 and A. Chmeiss and P. Jegou (1995) for PC-6. Unfortunately, we have remarked that PC-{5|6}, though it is widely better than PC-4 (Chmeiss and P. Jegou, 1996) was not very efficient in practice, especially for those classes of problems that require an important space to be run. So, we propose a new path-consistency algorithm called PC-8, the space complexity of which is O(n/sup 2/d) but its time complexity is O(n/sup 3/d/sup 4/), i.e. worse than that of PC-{5|6}. However, the simplicity of PC-8 as well as the data structures used for its implementation offer a higher performance than PC-{5|6}. The principle of PC-8 is also used to propose a new algorithm to achieve arc-consistency called AC-8.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"141 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124480192","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}
Introduces tensor-product neural networks, composed of a layer of univariate neurons followed by a net of polynomial post-processing. We look at the general approximation problem by these networks observing in particular their relationship to the Stone-Weierstrass theorem for uniform function algebras. The implementation of the post-processing as a two-layer network with logarithmic and exponential neurons leads to potentially important 'generalised' product networks which, however, require a complex approximation theory of the Mu/spl uml/ntz-Szasz-Ehrenpreis type. A backpropagation algorithm for product networks is presented and used in three computational experiments. In particular, approximation by a sigmoid product network is compared to that of a single-layer radial basis network and a multiple-layer sigmoid network.
{"title":"Perceptrons with polynomial post-processing","authors":"L. Sanzogni, Ringo Chan, R. Bonner","doi":"10.1109/TAI.1996.560792","DOIUrl":"https://doi.org/10.1109/TAI.1996.560792","url":null,"abstract":"Introduces tensor-product neural networks, composed of a layer of univariate neurons followed by a net of polynomial post-processing. We look at the general approximation problem by these networks observing in particular their relationship to the Stone-Weierstrass theorem for uniform function algebras. The implementation of the post-processing as a two-layer network with logarithmic and exponential neurons leads to potentially important 'generalised' product networks which, however, require a complex approximation theory of the Mu/spl uml/ntz-Szasz-Ehrenpreis type. A backpropagation algorithm for product networks is presented and used in three computational experiments. In particular, approximation by a sigmoid product network is compared to that of a single-layer radial basis network and a multiple-layer sigmoid network.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"106 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128319182","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}
Testing is a key issue in the design and production of digital circuits: the adoption of BIST (Built-in Self-Test) techniques is increasingly popular, but sometimes requires efficient algorithms for the automatic generation of the logic which generates the test vectors applied to the unit under test. This paper addresses the issue of identifying a cellular automaton able to generate input patterns to detect stuck-at faults inside a finite state machine (FSM). A suitable hardware structure is first identified. A genetic algorithm is then proposed, which directly identifies a cellular automaton able to reach a very good fault coverage of the stuck-at faults. The novelty of the method consists in combining the generation of test patterns with the synthesis of a cellular automaton able to reproduce them. Experimental results are provided, which show that in most of the standard benchmark circuits the cellular automaton selected by the genetic algorithm is able to reach a fault coverage close to the maximum one. Our approach is the first attempt of exploiting evolutionary techniques for identifying the hardware for input pattern generation in BIST structures.
{"title":"A genetic algorithm for automatic generation of test logic for digital circuits","authors":"Fulvio Corno, P. Prinetto, M. Reorda","doi":"10.1109/TAI.1996.560394","DOIUrl":"https://doi.org/10.1109/TAI.1996.560394","url":null,"abstract":"Testing is a key issue in the design and production of digital circuits: the adoption of BIST (Built-in Self-Test) techniques is increasingly popular, but sometimes requires efficient algorithms for the automatic generation of the logic which generates the test vectors applied to the unit under test. This paper addresses the issue of identifying a cellular automaton able to generate input patterns to detect stuck-at faults inside a finite state machine (FSM). A suitable hardware structure is first identified. A genetic algorithm is then proposed, which directly identifies a cellular automaton able to reach a very good fault coverage of the stuck-at faults. The novelty of the method consists in combining the generation of test patterns with the synthesis of a cellular automaton able to reproduce them. Experimental results are provided, which show that in most of the standard benchmark circuits the cellular automaton selected by the genetic algorithm is able to reach a fault coverage close to the maximum one. Our approach is the first attempt of exploiting evolutionary techniques for identifying the hardware for input pattern generation in BIST structures.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130299545","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}
M. Baker, L. Charnay, Michelle Joab, Benoît Lemaire, B. Safar, Daniel Schlienger
Analysis of expert medical diagnostic critiquing dialogues shows that explanation, argumentation and negotiation are strongly interlinked. After presenting analyses of a corpus of Wizard of Oz dialogues in this domain, we describe a design for a graphical interface that enables human-computer collaboration for the same task. Many of the dialogues' functionalities can be transferred to the interface, whilst avoiding natural language interpretation problems and providing a comparable degree of expressivity for the user.
{"title":"Incorporating functionalities of expert medical critiquing dialogues in the design of a graphical interface","authors":"M. Baker, L. Charnay, Michelle Joab, Benoît Lemaire, B. Safar, Daniel Schlienger","doi":"10.1109/TAI.1996.560442","DOIUrl":"https://doi.org/10.1109/TAI.1996.560442","url":null,"abstract":"Analysis of expert medical diagnostic critiquing dialogues shows that explanation, argumentation and negotiation are strongly interlinked. After presenting analyses of a corpus of Wizard of Oz dialogues in this domain, we describe a design for a graphical interface that enables human-computer collaboration for the same task. Many of the dialogues' functionalities can be transferred to the interface, whilst avoiding natural language interpretation problems and providing a comparable degree of expressivity for the user.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133408468","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}
After many studies, experiments, and assessments on the design and development of intelligent learning/teaching systems, we are now offering a production environment for these systems and shells. This environment is based on high abstraction level primitives: the 'NGE' kernel. Elaborated in a cognitive approach, these primitives have been defined on the notion of task (generic, cognitive, operating), not only to define the different levels of abstraction but also to improve the interactions between the universe of teachers, pedagogues and the universe of computer science.
{"title":"A task-based production environment for intelligent learning/teaching systems: the 'NGE' kernel","authors":"C. Canut, Murielle Eloi","doi":"10.1109/TAI.1996.560777","DOIUrl":"https://doi.org/10.1109/TAI.1996.560777","url":null,"abstract":"After many studies, experiments, and assessments on the design and development of intelligent learning/teaching systems, we are now offering a production environment for these systems and shells. This environment is based on high abstraction level primitives: the 'NGE' kernel. Elaborated in a cognitive approach, these primitives have been defined on the notion of task (generic, cognitive, operating), not only to define the different levels of abstraction but also to improve the interactions between the universe of teachers, pedagogues and the universe of computer science.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"237 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129332680","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 previously shown dynamical properties-dynamics of the activities for states-for higher order random neural networks, which use the weighted sum of products of input variables, with the digital state {1-,1} model. The paper describes dynamical properties for higher order random neural networks with the analog state models and the digital state (0,1) model.
{"title":"Dynamical properties of higher order random neural networks","authors":"H. Miyajima, Lixin Ma, Hiroyuki Suwa","doi":"10.1109/TAI.1996.560743","DOIUrl":"https://doi.org/10.1109/TAI.1996.560743","url":null,"abstract":"The authors have previously shown dynamical properties-dynamics of the activities for states-for higher order random neural networks, which use the weighted sum of products of input variables, with the digital state {1-,1} model. The paper describes dynamical properties for higher order random neural networks with the analog state models and the digital state (0,1) model.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"78 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125881815","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}
This paper focuses on the implementation of domain models for problem solving. We assume that the implementation should preserve the form of the domain models, and particularly the division between the knowledge structure and its role toward the solving process. We propose therefore to implement domain models with both knowledge bases and domain agents. We define a domain agent model which is linked to a knowledge base. We describe the various abilities of this agent and how it has been implemented in a multi-agent system with the YAFOOL language.
{"title":"Knowledge bases and agents for domain knowledge representation","authors":"M. Chouvet, F. Ber","doi":"10.1109/TAI.1996.560455","DOIUrl":"https://doi.org/10.1109/TAI.1996.560455","url":null,"abstract":"This paper focuses on the implementation of domain models for problem solving. We assume that the implementation should preserve the form of the domain models, and particularly the division between the knowledge structure and its role toward the solving process. We propose therefore to implement domain models with both knowledge bases and domain agents. We define a domain agent model which is linked to a knowledge base. We describe the various abilities of this agent and how it has been implemented in a multi-agent system with the YAFOOL language.","PeriodicalId":209171,"journal":{"name":"Proceedings Eighth IEEE International Conference on Tools with Artificial Intelligence","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124884531","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：