Exploratory Search in the Adaptive Social Semantic Web

Michal Tvarožek

Doctoral dissertation project supervised by prof. Mária Bieliková


Motivation and Goals

The Web has become a unique ubiquitous information source and a global socio-economic space used daily by roughly one third of the world’s population. Effective access to information on the Web via search, navigation and presentation of information thus presents key research challenges.

Typical problems that web users face daily include the navigation problem, where users lose track of their position in the web hyperspace and the direction they should take, short search queries and the resulting query ambiguity, irrelevance of search results and information overload. The end-result usually is that accessing information either takes ‘too long’ or that suitable information cannot be effectively accessed despite being available ‘somewhere’ in the Web.

Due to the critical nature of the aforementioned web usage issues for our society, a lot of research has been done on resolving them in different fields:

  • The Semantic Web envisioned by Tim Berners-Lee stresses the need for semantics and more robust querying systems using ontologies as common conceptualizations of knowledge.
  • The Adaptive Web movement lead by Peter Brusilovsky puts focus on personalization with respect to individual user needs and/or characteristics in order to improve all aspects of web-based systems.
  • The Social Web also often called Web 2.0 by O’Reilly, which focuses on interaction, user content contribution, information sharing and the human aspect of the Web.
  • Exploratory search, first coined by Gary Marchionini, which advocates changes in current end-user needs from finding of specific information to understanding information within its context, and to the successive processing / transformation of this information into new forms (e.g., an analysis, forecast).

Thus in our work, we focused on two primary areas of research in order to support effective end-user access to information:

  • Exploratory search with focus on end-user search and visualization support via multiple querying options and a comprehensive set of information visualization options.
  • User support during (semantic) query construction and modification via faceted browsing, semantic query generation and personalized recommendation.

Results

We devised a comprehensive exploratory search approach for the Semantic Web by extending and integrating partial approaches from the abovementioned fields. Our main results include the following:

  • Method for multi-paradigm exploration integrating keyword-based, view-based (faceted) and content-based search, navigation and visualization for semantic information spaces represented via ontologies. This enables us to provide end-users with advanced exploratory capabilities such as faceted query formulation, and task-centered result visualization and exploration.
  • Concept/item relevance model based on in-session user behavior and on short-/long-term user characteristics and social aspects considering user similarity and social relations. This allows us to infer the user relevance of items (e.g., facets, results) at run time and in real time.
  • Method for personalized recommendation of facets and restrictions in a faceted browser, using our underlying relevance model. Via personalization, we provide end-users with query construction and modification support, revisitation and orientation support and also reduce information overload and the navigation problem.
  • Method for dynamic facet generation from ontological metadata, which enables us to dynamically tailor the faceted querying and browsing interface to changes in the underlying information space accommodating for changes and updates (and even domain changes, assuming the required descriptive metadata are present in the ontology).
  • Method for adaptive view generation, which adapts the results overviews and result detail views based on our underlying relevance model (e.g., tailors attribute and/or column selection to individual user preferences). We provide adaptive table-based and matrix-based result overviews, and table-based and graph-based result details/relations views.

Conclusion

Our main contributions lie in the development of novel methods for navigation and presentation, and in the combination of approaches from the Semantic Web, Social Web and Adaptive Web initiatives. We devised a comprehensive faceted exploration approach for the Semantic Web and claim specific contribution to:

  • Multi-paradigm exploration – integrating view-based, content-based and keyword-based search with advanced adaptive visualizations and incremental graph exploration of both content and browsing history.
  • Personalized recommendation – devising a dynamic method of facet and restriction adaptation based on semantic logging of user action and continuous evaluation of the devised ontological user and relevance models.
  • Exploratory interface generation – devising a method for facet identification in ontological metadata, its transformation into interface widgets and their mapping onto the ontological querying backend (e.g., semantic search engines).

We thus improve upon the state of the art information exploration possibilities by providing end-users with effective means for browsing, presentation and understanding by incorporating semantics, adaptation, personalization and collaboration for seamless access to web resources, ultimately enabling end-user grade exploration of the Adaptive Social Semantic Web and achieving our original goals:

  • Empowering end-users with access to semantic information spaces by providing an end-user grade exploratory browser for the Semantic Web with interfaces for effective query formulation, result overview browsing and individual result exploration.
  • Facilitating the adoption of the Semantic Web by enabling Adaptive Social Semantic Web exploration for end-users via our exploratory search browser.

The thesis extended abstract is available in the Bulletin of the ACM Slovakia.

Selected publications

M. Tvarožek
Exploratory Search in the Adaptive Social Semantic Web. Dissertation thesis, Slovak University of Technology in Bratislava 2010 (in Slovak). pdf
M. Tvarožek, M. Bieliková
Bridging Semantic and Legacy Web Exploration: Orientation, Revisitation and Result Exploration Support. In WIRSS 2010: Proceedings of IEEE/WIC/ACM International Joint Conferences on Web Intelligence and Intelligent Agent Technologies, Workshop on Web Information Retrieval Support Systems (pp. 326–329). Toronto, Canada: IEEE CS (2010).
M. Tvarožek, M. Bieliková
Generating Exploratory Search Interfaces for the Semantic Web. In P. Forbrig, F. Patern?, & A. M. Pejtersen (Ed.), IFIP - World Computer Congress 2010: Human-Computer Interaction Symposium (pp. 175–186). Brisbane, Australia: Springer (2010).
M. Tvarožek, M. Bieliková
Factic: Personalized Exploratory Search in the Semantic Web. In B. Benatallah (Ed.), Lecture Notes in Computer Science, ICWE 2010: Web Engineering. Vol. 6189, pp. 528–531. Springer-Verlag, ISSN 0302-9763 (2010).
M. Tvarožek, M. Bieliková
Personalized Faceted Navigation in Semantically Enriched Information Spaces. In M. C. Angelides, P. Mylonas, & M. Wallace (Eds.), Advances in Semantic Media Adaptation and Personalization (Vol. 2, pp. 181–201). CRC Press (2009).
M. Tvarožek, M. Bieliková
Visualization of Personalized Faceted Browsing. In P. Forbrig, F. Patern?, & A. M. Pejtersen (Ed.), IFIP - World Computer Congress 2008: Human-Computer Interaction Symposium. IFIP 272, pp. 213–218. Milan, Italy: Springer (2008).

to Homepage to Teaching to the Top

Home
Research
Projects
Publications
Books
SCM
Teaching
Links
Last updated:
Mária Bieliková bielik [zavináč] fiit-dot-stuba-dot-sk
Design © 2oo1 KoXo