1,720,986 research outputs found
Learning Email Filtering Rules with Magi - A Mail Agent Interface
No full textAs the volume of data on the Internet increases the need for better tools to handle this flood of data is also growing. Interface agents are tools which are designed to aid the user in using various applications. This project describes the development of an agent which employs machine learning techniques to discover rules for filtering email. It explains how the agent observes the user in handling mail and how these observations are used to help automate this task. The agent is then evaluated, through testing, to examine whether such a tool can be useful as a personal assistant. A description of existing work is given, along with the design rationale, and a number of future extensions are suggested
Dimensionality Reduction and Representation for Nearest Neighbour Learning
No full textAn increasing number of intelligent information agents employ Nearest Neighbour learning algorithms to provide personalised assistance to the user. This assistance may be in the form of recognising or locating documents that the user might find relevant or interesting. To achieve this, documents must be mapped into a representation that can be presented to the learning algorithm. Simple heuristic techniques are generally used to identify relevant terms from the documents. These terms are then used to construct large, sparse training vectors. The work presented here investigates an alternative representation based on sets of terms, called set-valued attributes, and proposes a new family of Nearest Neighbour learning algorithms that utilise this set-based representation. The importance of discarding irrelevant terms from the documents is then addressed, and this is generalised to examine the behaviour of the Nearest Neighbour learning algorithm with high dimensional data sets containing such values. A variety of selection techniques used by other machine learning and information retrieval systems are presented, and empirically evaluated within the context of a Nearest Neighbour framework. The thesis concludes with a discussion of ways in which attribute selection and dimensionality reduction techniques may be used to improve the selection of relevant attributes, and thus increase the reliability and predictive accuracy of the Nearest Neighbour learning algorithm
Calendars, Schedules and the Semantic Web
No full textThe emergence of the Semantic Web has simplified and improved knowledge reuse across the Internet. By committing to various ontologies, an agent can now understand and reason about published information such as calendar events and schedules to meet its user's needs and provide assistance. We illustrate the benefits of garnering schedules from the Semantic Web for agent-based assistance, and introduce other initiatives being pursued by the RDF Calendar Taskforce
Agent Roles in Human Teams
No full textIn this paper, we describe results of a series of experiments investigating the effects of agent aiding on human teams. The role an agent played, its task, and the ease with which it communicated with its human teammates all influenced team behavior. Team supporting tasks such as relaying and reminding seemed particularly effective
Communicating Agents in Open Multi Agent Systems
No full textAgents often utilise the services of other agents to perform tasks within multi agent systems. To achieve this, an agent must first locate another agent that has the capability to provide a desired service (i.e. a service provider agent), and then interact with it. To communicate with a service provider, an agent requires information about: 1) the service provider agent's interface; 2) the ontology that defines concepts used by the provider agent; and 3) the agent communication language (ACL) the agent uses so that it can parse and understand the communication. Currently deployed MASs encode the interface description and the ontology within the capability description of a service provider, but assume a common ACL between communicating agents. Middle agents support the discovery of service providers based on the advertised capabilities of the service providers. This advertisement defines the provider agent's interface, and may reference the ontology used by the provider agent. However, the requester agent still requires information about the ACL used by the provider to be able to communicate with it. This paper demonstrates how agents can communicate with each other without making assumptions about the ACLs used, by presenting a template based shallow parsing approach to message construction/decomposition, thus greatly simplifying and improving the robustness of inter-agent communication
Task characteristics and Intelligent Aiding
No full textIn this paper, we describe the interactions between task characteristics and human agent interfaces in a team rendezvous route-planning task. The agents include an interface agent and two different task agents that perform similar tasks. The MokSAF interface agent links an Artificial Intelligence (AI) route planning agent to a Geographic Information System (GIS). Through this agent, the user specifies a start and an end point, and describes the composition and characteristics of a military platoon. Two aided conditions and one non-aided condition were examined. In the first aided condition, a route-planning agent (known as the Autonomous RPA) determines a minimum cost path between the specified end points. The user is allowed to define additional "intangible" constraints that describe situational or social information that should be considered when determining the route. In the second aided condition, a different agent, the Cooperative RPA, uses the same knowledge of the terrain and cost functions available to the Autonomous RPA, but restricts its search to paths within regions drawn by the user. In the unaided condition, Naive RPA, the user draws the route manually, then submits it to be tested against the terrain and cost functions for feasibility. Both aided conditions are superior to the control but differ in their relative effectiveness by scenario. In this paper we examine the varieties of challenges faced by commanders in two scenarios and relate them to the differential effectiveness of the agents
Learning Mechanisms for Information Filtering Agents
No full textIn recent years, software agents have been developed which assist users with tasks such as information filtering or information retrieval. Such systems have evolved from simple agents that refer to a user-defined script to filter incoming mail, to complex Web agents that not only learn their user's preferences but actively seek out Web pages that could be of interest. To provide personal assistance, an agent needs information about the user's interests and needs. This paper reviews how different mechanisms have been used to define a user profile, from simple rules to complex machine learning algorithms. Problems with user-defined scripts are discussed, as are the issues involved with integrating learning mechanisms into agents. One approach currently being developed to learn within an agent environment is then described
A formal model of the Semantic Web Service Ontology (WSMO)
Full text linkSemantic Web Service, one of the most significant research areas within the Semantic Web vision, has attracted increasing attention from both the research community and industry. The Web Service Modelling Ontology (WSMO) has been proposed as an enabling framework for the total/partial automation of the tasks (e.g., discovery, selection, composition, mediation, execution, monitoring and etc.) involved in both intra- and inter-enterprise integration of Web Services. To support the standardisation and tool support of WSMO, a formal model of the language is highly desirable. As several variants of WSMO have been proposed by the WSMO community, which are still under development, the syntax and semantics of WSMO should be formally defined to facilitate easy reuse and future development. In this paper, we present a formal Object-Z formal model of WSMO, where different aspects of the language have been precisely defined within one unified framework. This model not only provides a formal unambiguous model which can be used to develop tools and facilitate future development, but as demonstrated in this paper, can be used to identify and eliminate errors present in existing documentation
- …
