4-year full-time PhD research position in the WiBAF project

3/30463
label Burse autorenew 2025-09-29, 17:01 history_edu Silviu Marinescu
We offer a 4-year full-time PhD research position in the WiBAF project (Within-Browser Adaptation Framework) situated on the crossroads of the Adaptive Hypermedia, Data Mining and User Modeling areas. The project is funded by the Data Science research center of the Eindhoven University of Technology (TU/e), the Netherlands. The research will be conducted at the Information Systems Group, Department of Computer Science.

RESEARCH PROJECT AND TASK DESCRIPTION
Personalized information systems will undergo dramatic changes in the near future, caused in part by rising privacy concerns (and legislation) and in part by the need to distribute the computation involved in generating recommendations and performing adaptation.

In Web-based information systems data mining techniques are applied to model user intent, preferences or current context and thus facilitate data-driven personalization. Such data-driven user modeling has been heavily used already in recommender systems that employ content-based and collaborative filtering, and in search engines with personalized or context-aware ranking of organic search results or ad placement. In the area of personalization and adaptation of content, interaction and navigation, with applications in e.g. e-learning and e-culture, either hand-crafted adaptation rules or semantically rich databases have been used as the basis for deciding how to perform personalization.



The main objective for the PhD student in WiBAF project is to investigate how to bring these two currently distinct and to a large extent isolated approaches to personalization, i.e. purely hand-crafted and purely data-driven, together to perform within-site or even within-page adaptation based on both the analysis of user behavior and background knowledge of a certain domain.

One of the concrete targets is defining, studying and implementing a complete framework including a generic language for expressing adaptation and for generating client-side code that performs this within-browser adaptation, based on local client-side user model and global adaptation knowledge available at the server-side.

The framework should enable meta-adaptation which continuously adapts (updates) the adaptation process as user behavior changes. This makes use of research into the topics of explore-exploit trade-off in adaptive information systems like recommenders, context-awareness and recurrent concept drift in data mining.

It is expected that the PhD candidate will be involved in the complete R&D cycles including theory, framework and techniques development; implementation of the software prototypes; and experimentation with real use cases that are facilitated by the industrial collaborator - Digital Innovations Lab of the Web analytics company Adversitement B.V.