A Computational Framework for Multi-dimensional Context-aware Adaptation
Download as PPTX, PDF1 like302 views
This document proposes a computational framework for multi-dimensional context-aware adaptation. It aims to transform different aspects of a system according to context to provide high usability. Current approaches are often limited to single contexts or platforms. The proposed framework would consider multiple contexts, dimensions, and levels of an application to support adaptation. It involves systematic reviews of adaptation concepts, UML modeling of context information, an algorithms library, and machine learning techniques to provide context-aware adaptation with evaluation of usability. The goal is to develop a unified approach for context-aware adaptation across contexts, dimensions, and levels.
![A Computational Framework for Multi-dimensional Context-aware AdaptationVivian Genaro MottiLouvain Interaction Laboratory (LILAB)Universitécatholique de Louvain (UCL)Vivian.genaromotti@uclouvain.beIntroduction Adaptation consists in transforming, according to the context, different aspects of a system, in different levels, in order to provide users an interaction of high usability levelMotivation Most of the applications are often developed considering a pre-defined context of use, however, not only the contexts of use and users are heterogeneous, but users also interact with applications via different devices, platforms and meansChallenges and Shortcomings Consider all context information to provide users adaptation with high usability level and transparency The works reported so far are often limited to one dimension orplatform at a time; the current approaches are not unified, inconsistencies, e.g. in terminology, are commonGoal Develop a framework to support the implementation of adaptation considering different contexts of use, dimensions and levels of an application subject to adaptation, aiming a high usability levelMethodology A Systematic Review to gather adaptation concepts (techniques, strategies, approaches, and models) A template to define adaptation techniques regarding content (audio, image, text), presentation and navigation UML diagrams to model the context information An Algorithms Library to implement adaptation techniques Advanced Logic Algorithms using Machine Learning techniques to provide context-aware adaptation (e.g. Decision Tree, Bayesian Network and Hidden Markov Model) Iterative usability evaluations Case studies to verify the feasibilityResults A systematic review is being performed continuously: 89 techniques were documented with templates, detailed, analyzed and compared Models are being created in UML to model the context (Use Case, Class Diagram, State Machine, Sequence Diagram) An Algorithms Library is being developed with the techniques gathered Machine learning algorithms are being investigated to combine information and provide adaptationFuture Work Implement the machine learning techniques Define precisely the evaluation plan, perform evaluation Perform the case studiesFinal Remarks It is a challenge to provide users adaptation without disturbing and confusing them, user evaluation is, then, necessary to achieve a higher level of usability. A wide approach is necessary to cover and try to unify the current knowledge about context-aware adaptation.feedback[Serenoa]eSerenoa is aimed at developing a novel, open platform for enabling the creation of context-sensitive SFE. TitleTextImageSerenoaLorem ipsum lorem ipsum lorem ipsum loremipsum](https://image.slidesharecdn.com/genaro-eics2011-poster-110831093934-phpapp02/85/A-Computational-Framework-for-Multi-dimensional-Context-aware-Adaptation-1-320.jpg)
A Computational Framework for Multi-dimensional Context-aware Adaptation
- 1. A Computational Framework for Multi-dimensional Context-aware AdaptationVivian Genaro MottiLouvain Interaction Laboratory (LILAB)Universitécatholique de Louvain (UCL)Vivian.genaromotti@uclouvain.beIntroduction Adaptation consists in transforming, according to the context, different aspects of a system, in different levels, in order to provide users an interaction of high usability levelMotivation Most of the applications are often developed considering a pre-defined context of use, however, not only the contexts of use and users are heterogeneous, but users also interact with applications via different devices, platforms and meansChallenges and Shortcomings Consider all context information to provide users adaptation with high usability level and transparency The works reported so far are often limited to one dimension orplatform at a time; the current approaches are not unified, inconsistencies, e.g. in terminology, are commonGoal Develop a framework to support the implementation of adaptation considering different contexts of use, dimensions and levels of an application subject to adaptation, aiming a high usability levelMethodology A Systematic Review to gather adaptation concepts (techniques, strategies, approaches, and models) A template to define adaptation techniques regarding content (audio, image, text), presentation and navigation UML diagrams to model the context information An Algorithms Library to implement adaptation techniques Advanced Logic Algorithms using Machine Learning techniques to provide context-aware adaptation (e.g. Decision Tree, Bayesian Network and Hidden Markov Model) Iterative usability evaluations Case studies to verify the feasibilityResults A systematic review is being performed continuously: 89 techniques were documented with templates, detailed, analyzed and compared Models are being created in UML to model the context (Use Case, Class Diagram, State Machine, Sequence Diagram) An Algorithms Library is being developed with the techniques gathered Machine learning algorithms are being investigated to combine information and provide adaptationFuture Work Implement the machine learning techniques Define precisely the evaluation plan, perform evaluation Perform the case studiesFinal Remarks It is a challenge to provide users adaptation without disturbing and confusing them, user evaluation is, then, necessary to achieve a higher level of usability. A wide approach is necessary to cover and try to unify the current knowledge about context-aware adaptation.feedback[Serenoa]eSerenoa is aimed at developing a novel, open platform for enabling the creation of context-sensitive SFE. TitleTextImageSerenoaLorem ipsum lorem ipsum lorem ipsum loremipsum