Task Force on User Adaptive Search Interfaces

Current Events

This task force has been formed as part for the European Network of Excellence on Intelligent Technologies for Smart Adaptive Systems (EUNITE) to review the state of the art in user adaptive search interfaces and to initiate and intensify research collaboration between different research communities and the industry.

This page provides an overview of the project goals and the activity of the task force. If you have any question regarding the task force or like to join send an email to Andreas N�rnberger.

Motivation and brief task description

Over the last years a number of powerful methods have been proposed and realized, that enable a user to search in or browse large collections of documents which may consist of web pages, text or multimedia documents. Most of these methods provide either a simple keyword search interface (e.g., search engines like AltaVista and Google) or a fixed hierarchy or index for navigation (e.g., almost all library systems or the web search engine Yahoo). Unfortunately, the hierarchies and indexes are frequently subjectively labeled and often disjunctive. Therefore, on each level of the hierarchy a user has to select a specific keyword describing a category, even if the searched document might have been assigned to more than one listed category. This is especially critical if the user would assign a document to a different category and therefore might not find the searched information. A main problem of keyword search systems is that - even though the provided result list usually includes the searched documents if the keywords are well chosen - the documents in the result list are only sorted with respect to the given search terms (and usually some further criteria derived from the document collection). Documents belonging to different domains appear rather unsorted in the list. Meanwhile, search engines like Teoma or WiseNut try to classify the search results. However, similar to fixed hierarchies, user specific interests are neglected.

The aim of this task force is to collect state-of-the art methods that enable the use of user interests in the design of search interfaces. This should include methods for modeling user interests, analyzing user behavior (profiling) as well as design aspects of the interface itself (visualization or other techniques to support a user in browsing a document collection). Furthermore, a simple prototypical implementation should be developed and provided to the community.

Past Activities

As part of the activities of the task force, the following Workshops have been organized:

Participants

The scholars and affiliated research groups listed in the following are currently involved in the activity of the task force:

References

Books

Articles in Conference and Workshop Proceedings

Otto-von-Guericke University Magdeburg
School of Computer Science
Institute of Knowledge Processing and Language Engineering
Information Retrieval Group

� 2003, Andreas N�rnberger