Utilizing Context History to Provide Dynamic Adaptations.

Cheverst, Keith and Byun, Hee Eon (2004) Utilizing Context History to Provide Dynamic Adaptations. Applied Artificial Intelligence, 18 (6). pp. 533-548. ISSN 0883-9514

Full text not available from this repository.

Abstract

In order to provide "intimate" and "dynamic" adaptations under Weiser's vision for ubiquitous computing environments, we propose the utilization of context history together with user modeling and machine learning techniques. Our approach supports proactive adaptations by inducing patterns of user behavior. In addition, we support the requirement for enabling the user to receive an explicit and understandable explanation when a proactive adaptation occurs in order to encourage a trust relationship between the user and the context-aware system. In this article, we describe an experiment to examine the feasibility of our approach for supporting proactive adaptations in the domain of an intelligent office environment. The initial results of our experiment are promising and demonstrate how our system could gradually learn the user's preferences for controling his office environment by making inductions from the context history. Based on these initial findings, we believe that context history has a concrete role to play in supporting proactive adaptation in a ubiquitous computing environment.

Item Type:
Journal Article
Journal or Publication Title:
Applied Artificial Intelligence
Additional Information:
The paper describes the highly original and significant AI approach used to support a 'proactive' intelligent office system. The approach utilises fuzzy decision trees to support the inference of explainable rules from training data in the form of context history. The decision tree approach was motivated by an HCI informed understanding of the need to ensure user control and associated user understanding when developing proactive systems. The journal's readership includes ""Researchers involved with artificial intelligence, robotics, and computers; R & D personnel; consultants; AI hardware/software vendors; and academics in the fields of AI, computer science, CAD, CAM, and management engineering'. RAE_import_type : Journal article RAE_uoa_type : Computer Science and Informatics
Uncontrolled Keywords:
/dk/atira/pure/researchoutput/libraryofcongress/qa75
Subjects:
?? ARTIFICIAL INTELLIGENCEQA75 ELECTRONIC COMPUTERS. COMPUTER SCIENCE ??
ID Code:
2593
Deposited By:
Deposited On:
28 Mar 2008 12:52
Refereed?:
Yes
Published?:
Published
Last Modified:
17 Sep 2023 00:30