Course on Intelligent Interfaces (INT5, 3p)

Below follows a couple of examples of Masters work in the area of Intelligent Interfaces. We are most willing to discuss other ideas from students who take the course.

Information Filtering on the www

The www is growing fast and inexperienced users have a hard time filtering and searching for information relevant to their needs. Since www is in most parts unstructured and growing, any aid in filtering will have to be adaptive to the dynamic changes of the information space.

One approach to aiding a user to filter for information is to base the filtering on cooperative user modeling: i.e. the modelling builds upon a group of users and their preferences. An individual user can ask to see the same information as a group of users that the user share interests with.

This Masters work will be in exploring machine learning as a tool for extracting user group preferences. It will also involve designing an interface that enables a user to decide which group of users they share preferences with.

Some previous work by Patti Maes, Jon Orwant, and others, will be used as the basis for the work.

Intelligent Multimodal Navigation Support

Combining media to present information is a difficult task. The task of the system producing the information must be modelled in detail, the users preferences and cognitive abilities must be well-understood, etc.

This Masters work will be to explore different media and their potential usage as a means for helping users to navigate in hypermedia structures. Previous work has been to offer fish-eye views, dialogue histories, etc. Building upon that work and adding ideas of visual momentum the work will be centrered around designing a tool for an intelligent multimodal navigation help system.

E-mail kia@sics.se, URL http://www.sics.se/~kia

E-mail annika@sics.se, URL http://www.sics.se/~annika