Fuzzy neural networks are connectionist systems that facilitate learning from data, reasoning over fuzzy rules, rule insertion, rule extraction, and rule adaptation. The concept evolving fuzzy neural networks (EFuNNs), with respective algorithms for learning, aggregation, rule insertion, rule extraction, is further developed here and applied for on-line knowledge discovery on both prediction and classification tasks. EFuNNs operate in an on-line mode and learn incrementally through locally tuned elements. They grow as data arrive, and regularly shrink through pruning of nodes, or through node aggregation. The aggregation procedure is functionally equivalent to knowledge abstraction. The features of EFuNNs are illustrated on two real-world application problems—one from macroeconomics and another from Bioinformatics. EFuNNs are suitable for fast learning of on-line incoming data (e.g., financial and economic time series, biological process control), adaptive learning of speech and video data, incremental learning and knowledge discovery from growing databases (e.g. in Bioinformatics), on-line tracing of processes over time, life-long learning. The paper includes also a short review of the most common types of rules used in the knowledge-based neural networks for knowledge discovery and data mining.
+Fuzzy neural networks are connectionist systems that facilitate learning from data, reasoning over fuzzy rules, rule insertion, rule extraction, and rule adaptation. The concept evolving fuzzy neural networks (EFuNNs), with respective algorithms for learning, aggregation, rule insertion, rule extraction, is further developed here and applied for on-line knowledge discovery on both prediction and classification tasks. EFuNNs operate in an on-line mode and learn incrementally through locally tuned elements. They grow as data arrive, and regularly shrink through pruning of nodes, or through node aggregation. The aggregation procedure is functionally equivalent to knowledge abstraction. The features of EFuNNs are illustrated on two real-world application problems—one from macroeconomics and another from Bioinformatics. EFuNNs are suitable for fast learning of on-line incoming data (e.g., financial and economic time series, biological process control), adaptive learning of speech and video data, incremental learning and knowledge discovery from growing databases (e.g. in Bioinformatics), on-line tracing of processes over time, life-long learning. The paper includes also a short review of the most common types of rules used in the knowledge-based neural networks for knowledge discovery and data mining.
Keywords: fuzzy rules evolving fuzzy neural networks @@ -14,19 +14,19 @@ macroeconomics bioinformatics
-Download (PDF, 321 KB)
+Download (PDF, 707 KB)
Agent-oriented software engineering is a promising new approach to software engineering that uses the notion of an agent as the primary entity of abstraction. The development of methodologies for agent-oriented software engineering is an area that is currently receiving much attention, there have been several agent-oriented methodologies proposed recently and survey papers are starting to appear. However the authors feel that there is still much work necessary in this area; current methodologies can be improved upon. This paper presents a new methodology, the Styx Agent Methodology, which guides the development of collaborative agent systems from the analysis phase through to system implementation and maintenance. A distinguishing feature of Styx is that it covers a wider range of software development life-cycle activities than do other recently proposed agent-oriented methodologies. The key areas covered by this methodology are the specification of communication concepts, inter-agent communication and each agent’s behaviour activation—but it does not address the development of application-specific parts of a system. It will be supported by a software tool which is currently in development.
+Agent-oriented software engineering is a promising new approach to software engineering that uses the notion of an agent as the primary entity of abstraction. The development of methodologies for agent-oriented software engineering is an area that is currently receiving much attention, there have been several agent-oriented methodologies proposed recently and survey papers are starting to appear. However the authors feel that there is still much work necessary in this area; current methodologies can be improved upon. This paper presents a new methodology, the Styx Agent Methodology, which guides the development of collaborative agent systems from the analysis phase through to system implementation and maintenance. A distinguishing feature of Styx is that it covers a wider range of software development life-cycle activities than do other recently proposed agent-oriented methodologies. The key areas covered by this methodology are the specification of communication concepts, inter-agent communication and each agent’s behaviour activation—but it does not address the development of application-specific parts of a system. It will be supported by a software tool which is currently in development.
Keywords: agent-based software engineering methodologies for agent-oriented software development
-Download (PDF, 103 KB)
+Download (PDF, 153 KB)
Download (PDF, 451 KB)
+Download (PDF, 488 KB)
This paper discusses technology to support the use of UML for representing ontologies and domain knowledge in the Semantic Web. Two mappings have been defined and implemented using XSLT to produce Java classes and an RDF schema from an ontology represented as a UML class diagram and encoded using XMI. A Java application can encode domain knowledge as an object diagram realised as a network of instances of the generated classes. Support is provided for marshalling and unmarshalling this object-oriented knowledge to and from an RDF/XML serialisation.
-Download (PDF, 912 KB)
+Download (PDF, 482 KB)
Download (PDF, 131 KB)
+Download (PDF, 216 KB)
Download (PDF, 155 KB)
+Download (PDF, 293 KB)
Ontologies play an important role in defining the terminology that agents use in the exchange of knowledge-level messages. As object-oriented modelling, and the Unified Modeling Language (UML) in particular, have built up a huge following in the field of software engineering and are widely supported by robust commercial tools, the use of UML for ontology representation in agent systems would help to hasten the uptake of agent-based systems concepts into industry. This paper examines the potential for UML to be used for ontology modelling, compares it to traditional description logic formalisms and discusses some further possibilities for applying UML-based technologies to agent communication systems.
-Download (PDF, 102 KB)
+Download (PDF, 143 KB)
This paper examines a recent trend amongst software agent application and platform developers to desire the ability to send domain-specific objects within inter-agent messages. If this feature is to be supported without departing from the notion that agents communicate in terms of knowledge, it is important that the meaning of such objects be well understood. Using an object-oriented metamodelling approach, the relationships between ontologies and agent communication and content languages in FIPA-style agent systems are examined. It is shown how object structures in messages can be considered as expressions in ontology-specific extensions of standard content languages. It is also argued that ontologies must distingish between objects with and objects without identity. Traditionally ontologies are used in agent systems “by reference”. An agent is not required to explicitly reason with the ontology, or even to have an online copy available. The names of ontologies can simply be used as a contract between agents undertaking a dialogue: they each claim to be using an interpretation of the terms used in the conversation that conforms to the ontology. The content language uses a string-based syntax to represent sentences in the language which are constructed using constants and function and predicate symbols from the ontology as well as built-in language symbols such as “and” and “or”.
+This paper examines a recent trend amongst software agent application and platform developers to desire the ability to send domain-specific objects within inter-agent messages. If this feature is to be supported without departing from the notion that agents communicate in terms of knowledge, it is important that the meaning of such objects be well understood. Using an object-oriented metamodelling approach, the relationships between ontologies and agent communication and content languages in FIPA-style agent systems are examined. It is shown how object structures in messages can be considered as expressions in ontology-specific extensions of standard content languages. It is also argued that ontologies must distingish between objects with and objects without identity. Traditionally ontologies are used in agent systems “by reference”. An agent is not required to explicitly reason with the ontology, or even to have an online copy available. The names of ontologies can simply be used as a contract between agents undertaking a dialogue: they each claim to be using an interpretation of the terms used in the conversation that conforms to the ontology. The content language uses a string-based syntax to represent sentences in the language which are constructed using constants and function and predicate symbols from the ontology as well as built-in language symbols such as “and” and “or”.
-Download (PDF, 131KB)
+Download (PDF, 171 KB)
Download (PDF, 207 KB)
+Download (PDF, 139 KB)
It focuses on requirements and applications, and is intended to provide a real-world, New Zealand-oriented context for the research in distributed information technologies (DIST).
-The report represents the culmination of a series of workshops, industrial consultations, a questionnaire, and the experiences of the authors’ institutions during the project, and therefore it supplements any previously produced material.
+The report represents the culmination of a series of workshops, industrial consultations, a questionnaire, and the experiences of the authors’ institutions during the project, and therefore it supplements any previously produced material.
-Download (PDF, 1.8 MB)
+Download (PDF, 1.3 MB)
This paper reports on a field study into the nature of decision making in the command and control of emergency ambulances at the London Ambulance Service (LAS). This paper will describe how real-time decisions are made by emergency medical dispatchers and the decision strategies they invoke as they assess the situation, plan and co-ordinate the dispatch of emergency ambulances.
-A cognitive task analysis approach known as the Critical Decision Method was used in the study. The study showed that decision making in emergency ambulance command and control involves four major processes—assessment of the situation, assessment of resources, planning, and co-ordinating and control. These four processes function within an awareness of goings-on in and around the sectors that the dispatchers operate in. This awareness is referred to as situation awareness and is being reported elsewhere. The decision making process resembles the decision making described by naturalistic decision making models and is an extension of the Integrated Decision Model. The study also suggested that a lot of effort was directed at understanding and assessing the situation and in maintaining a constant awareness of the situation. These observations have significant implications for the design of information systems for command and control purposes. These implications will be discussed separately in another paper.
+A cognitive task analysis approach known as the Critical Decision Method (Hoffman et al., 1998; Klein et al., 1989) was used in the study. The study showed that decision making in emergency ambulance command and control involves four major processes—assessment of the situation, assessment of resources, planning, and co-ordinating and control. These four processes function within an awareness of goings-on in and around the sectors that the dispatchers operate in. This awareness is referred to as situation awareness and is being reported elsewhere (Wong & Blandford, submitted). The decision making process resembles the decision making described by naturalistic decision making models (see (Zsambok & Klein, 1997) for an extensive discussion on the topic) and is an extension of the Integrated Decision Model (Wong, 1999). The study also suggested that a lot of effort was directed at understanding and assessing the situation and in maintaining a constant awareness of the situation. These observations have significant implications for the design of information systems for command and control purposes. These implications will be discussed separately in another paper.
The paper will first introduce the domain of EMD at the LAS, then explain how the Critical Decision Method was used in the data collection and in the data anlaysis. It will then describe how decisions are made, particularly during major incidents, and then discuss the implications of those findings for the design of command and control systems.
-Download (PDF, 262 KB)
+ +Download (PDF, 204 KB)