Personalized Interaction in Web Applications Based on Semantic Web

Document Sample
Personalized Interaction in Web Applications Based on Semantic Web Powered By Docstoc
					TU e

technische universiteit eindhoven

Personalized Interaction in Web Applications Based on Semantic Web Technology
Passepartout (ITEA ip04017)  Coupling home media-centers to home
networks for rendering scalable content from High Definition television (HD-TV) to lower definitions is a seamless fashion. This includes making a step towards ambient intelligence through mass personalization of reactive content (Figure 1).
- Consumer s
- Aggregated - Personalised

Information Systems Group user can behave very differently given the current  The information systems group at the TU/e context. has expertise in designing and building  Systems have difficulties to accommodate to take
software systems for storage and distribution of information.  One of the current research areas is Personalized Interaction in Web Applications based on Semantic Web Technology.  The research is distributed over a number of related external and internal projects.  Here we present a selection of these projects. very different view-point (e.g. the ability to accommodate both domain-experts and novices)  User communication (i.e. for data validation and knowledge probing) is usually very limited
- Value Chain
- Past - Future

- Content
- Passive - Reactive

- Network
- Separate - Converged

- Devices
- Standalone - Networked

 TU/e contributes in the field of
content filtering and adaptation  A first task is to personalize the display and interactivity of the system. Then create adaptive assistants and recommendation programs with support of interactive user modeling.

Issues of current Personalized systems  Applications cannot exchange data on the
user.  User modeling is often very application dependent.  Current systems often do not regard that the.

- Present

- Prosumer creates own content, but enjoys only some interactivity outside PC - Users and communities using reactive-interactive principles

- Internet and other networks divided and poorly managed

- Connected digitally, but no network management possible

- Only passive (TV) and active (PC) modes exist: Proto-Timeshifters arrive - Consumers have control of content in form and presentation

- Future

- Fibre, Blu-Ray, UWB - Media center as and WiMax networks staging areas for dominant providers to feed multiple in-home devices

Figure 1. Towards reactive content that can be accessed with personalized home networks.

SWALE (NWO) types of users  Identify dimensions of knowledge sharing that  Focus on knowledge representation problems
can be applied to learning systems to enable different viewpoints on learning material.  Facilitate support joint construction of educational materials and adaptation to individual teacher/learner needs.  Formalize the construction of shared knowledge to provide effective author guidance and learner feedback in environments for individualized learning and course delivery.  Based on the earlier OntoAims (Figure 2) and Style projects. introduced by multimedia aspects.  Creation of ontological descriptions of the domain which can be revealed in different levels of detail to different users.  A first architecture (Figure 3) and demos are available.

 TU/e focuses on context-based profiling for applications.  This includes issues like context-determination, data distribution and semantic interoperability.  Figure 4 depicts a high level architecture of the MobiLife profile management component.
«interface» ContextProvider «interface» ContextConsumer

MobiLife (IST-511607)  Bring advances in mobile applications and services
within the reach of users in their everyday life.  Combine Personalization, context-awareness, privacy and trust, as well as multi-modal interfaces to achieve the best possible user experience.
User Modeling Broker
matched UM fragments Registered Applications query call services Utility Services

«interface» ProfileEditor

«interface» IPMContextProvider

«interface» IPMContextConsumer

«interface» IPMPrivacyAndTrust

CHIME (NWO)
 Investigate the use of semantic models for tailoring the presentation of cultural information extracted from existing repositories to different
Concept & Content Browser
Learning Resource Library
uses

SearchAlgorithm

ProfileManager

Profile
1 1 1 1..* 1

add integrated data query current UM

Resource Model

Interactive User Modelling
Activity User Profile OWL-OLM Dialog

AHA!

Publisher Matchmaker Ontology Repository
Entry
1..* 1

uses

Course Task Model Course Task Sequencing
uses uses & updates

User Model of AHA!

DistributionAbstractionLayer
1 1..*

SampLe

Reasoner Invoker

1..* ProfileSubSet

User Model Domain Ontology Browser
uses uses

User Model of SampLe

Domain Ontology

Figure 2. OntoAims integrated architecture

Figure 3. Chime Architecture

Figure 4. MobiLife Profile Management Architecture

HERA methodology (internal project)  Methodology for the design of personalized
Adaptive Web Information Systems (AWIS).  Create specialized AWIS by configuration by means of specific models.

 Simplify design of an AWIS by dividing the
process into three logical layers (content, navigation and presentation).  Includes extensive support for context and personalization  Provides the primitives to establish the complete datasource (e.g. an application) user interaction.

Future Work  Progress with the introduced projects.  Combine and integrate the results of the
project in an extension to HPG enabling these technologies.

HPG

 Toolset that implements the HERA
methodology.  Both in client-side (Figure 5) and server-side “flavor”.  Graphical (Visio-based) tool support for constructing the needed models.  Figure 6 is an example of an HPG output adapted to both a normal web browser and a Mobile phone.
Figure 5. Client-side HPG user interface Figure 6. Generated presentations: HTML and WML

Information Systems Group – http://wwwis.win.tue.nl/~hera – hera@win.tue.nl – +31 40 247 27 33

/ department of mathematics and computer science