-
[show abstract]
[hide abstract]
ABSTRACT: Associative memory models have been developed as mathematical and cognitive models to represent human memory. The mathematical models have been successfully applied in classification and pattern recognition tasks. But there is a practical value in the cognitive representation of it which has given rise to associative memory models known as episodic associative memory (EAM). To date, the practical application of EAMs is less common than the mathematical models mainly due to representation deficiencies in EAMs. This paper proposes a novel approach known as episodic associative memory with a neighborhood effect (EAMwNE), which overcomes these representative limitations and increases the practical value of EAMs. EAMwNE utilises the lateral relationships among feature values, considering a neighborhood in achieving this. Consequently, each data pattern is represented by a number of memory traces (opposed to the single representation in early models) depending on the size of neighborhood. In this paper, EAMwNE is experimented with classification tasks to demonstrate the enhanced capabilities of it. The accuracy of EAMwNE is further enhanced by a 'reward' mechanism.
TENCON 2005 2005 IEEE Region 10; 12/2005
-
[show abstract]
[hide abstract]
ABSTRACT: Distributed data mining (DDM) performs partial analysis of data at distributed locations and sends a summarized version to the peer sites or a central location for further analysis. Meta-learning is a technique that generates local classifiers (concepts or models) from distributed data sets to use in producing a global classifier. This inherently distributed nature of meta-learning provides much advantage in implementing practical DDM systems. Currently machine learning techniques such as supervised neural networks, decision trees, rules and genetic algorithms are used in the meta-learning process. Inspired by the cognitive representation of human memory, this paper presents a novel mechanism known as concept-episodic associative memory with a neighborhood effect (C-EAMwNE) to compute meta-classifiers. C-EAMwNE is an enhanced version of EAMwNE model previously developed by the authors which overcomes practical limitations of other existing cognitive representations. C-EAMwNE is applied to a multi-agent DDM system with learning agents and a central administrator agent. Learning agents use C-EAMwNE to generate meta-classifiers at distributed data sites and communicate them to the central administrator agent (CAA). CAA produces a final concept description from the distributed classifiers to be used in classification tasks.
Intelligent Agent Technology, IEEE/WIC/ACM International Conference on; 10/2005
-
[show abstract]
[hide abstract]
ABSTRACT: Distributed Data Mining (DDM) performs partial analysis of data at distributed locations and sends a summarized version to the peer sites or a central location for further analysis. Meta-learning is a technique that generates local classifiers (concepts or models) from distributed data sets to use in producing a global classifier. This inherently distributed nature of metalearning provides much advantage in implementing practical DDM systems. Currently machine learning techniques such as supervised neural networks, decision trees, rules and genetic algorithms are used in the metalearning process. Inspired by the cognitive representation of human memory, this paper presents a novel mechanism known as Concept-Episodic Associative Memory with a Neighborhood Effect (C-EAMwNE) to compute metaclassifiers. C-EAMwNE is an enhanced version of EAMwNE model previously developed by the authors which overcomes practical limitations of other existing cognitive representations. C-EAMwNE is applied to a multi-agent DDM system with learning agents and a central administrator agent. Learning agents use CEAMwNE to generate meta-classifiers at distributed data sites and communicate them to the central administrator agent (CAA). CAA produces a final concept description from the distributed classifiers to be used in classification tasks.
Intelligent Agent Technology, IEEE / WIC / ACM International Conference on. 09/2005;
-
[show abstract]
[hide abstract]
ABSTRACT: Intelligent agent technology, which expects to combine the marked trends in history of computing such as ubiquity, interconnection, intelligence, delegation and human orientation can be considered as a step towards the next stage of artificial intelligence. This new technology attempts to reduce the gap between man and machine. The remarkable ability of a human being to make decisions is art ongoing learning and evolutionary process. Therefore, when reducing the man-machine gap, one of the main issues to address is how to make agents decision makers in a human oriented way. The paper presents novel conceptual agent framework to provide human like decisions inspired by human behavior and brain study models. The proposed learning and evolutionary agent architecture make the agent capable of handling the dynamism in the environment too. The experiments illustrated with the banking application demonstrate how the proposed framework enables a software agent to make decisions in a human oriented manner.
Hybrid Intelligent Systems, 2004. HIS '04. Fourth International Conference on; 01/2005
-
[show abstract]
[hide abstract]
ABSTRACT: Intelligent agent technology can be considered as a step towards the next stage of artificial intelligence. This new technology attempts to bridge the gap between man and machine. When bridging the man-machine gap, one of the main issues to address is how to make agents capable of autonomous actions in a dynamic environment. Inspired by human behavior, psychology and brain science, This work presents a layered agent architecture which combines two fundamental forms of adaptation: learning and evolution. Each layer depicts the functionality of a human being making the agent better prepared to face environment dynamisms. The novel feature of the proposed architecture is, it enables the agent to evolve such that, the best action required for a given state of the environment is identified through learning rather than using a pre defined set of actions or plans.
Intelligent Agent Technology, 2004. (IAT 2004). Proceedings. IEEE/WIC/ACM International Conference on; 10/2004
-
[show abstract]
[hide abstract]
ABSTRACT: This paper describes a system, which has the capability to analyze and discover knowledge gathered from distributed sources. This autonomous system is implemented using the growing self organizing map (GSOM), which is a dynamic version of the self organizing map (SOM) and a multi agent communication system. The important contribution of this model is its ability to transfer local knowledge to a central knowledge base referred to as a central controller where the data is clustered and analyzed to identify similarities and variations. The central controller has the capability of transferring back its global knowledge to the distributed sources.
Intelligent Agent Technology, 2003. IAT 2003. IEEE/WIC International Conference on; 11/2003
-
[show abstract]
[hide abstract]
ABSTRACT: Supervised and unsupervised self generating neural network
architectures have been used in the recent past. Our previous work
(1998) has described an unsupervised self generating feature map, called
the growing self organising map (GSOM). In this paper we describe some
extensions to the GSOM such that it could be used to map and analyse
more realistic data sets
Neural Networks, 1999. IJCNN '99. International Joint Conference on; 02/1999
-
[show abstract]
[hide abstract]
ABSTRACT: This paper describes a system, which has the capability to analyze and discover knowledge gathered from distributed sources. This autonomous system is implemented using the Growing Self Organizing Map (GSOM), which is a dynamic version of the Self Organizing Map (SOM) and a multi agent communication system. The important contribution of this model is its ability to transfer local knowledge to a central knowledge base referred to as a central controller where the data is clustered and analyzed to identify similarities and variations. Central controller has the capability of transferring back its global knowledge to the distributed sources.
Intelligent Agent Technology, IEEE / WIC / ACM International Conference on.
-
[show abstract]
[hide abstract]
ABSTRACT: Intelligent agent technology can be considered as a step towards the next stage of artificial intelligence. This 7new technology attempts to bridge the gap between man and machine. When bridging the man-machine gap, one of the main issues to address is how to make agents capable of autonomous actions in a dynamic environment. Inspired by human behavior, psychology and brain science, this paper presents a layered agent architecture which combines two fundamental forms of adaptation: learning and evolution. Each layer depicts the functionality of a human being making the agent better prepared to face environment dynamisms. The novel feature of the proposed architecture is, it enables the agent to evolve such that, the best action required for a given state of the environment is identified through learning rather than using a pre defined set of actions or plans.
Intelligent Agent Technology, IEEE / WIC / ACM International Conference on.
