Intelligent Multimedia Interface Agents

Wolfgang Wahlster and Elisabeth Andre

Course Description

The goal of this tutorial is to survey a new generation of highly personalized agent-based or assistant-like user interfaces. Personalization refers to the ability of an interface to adapt its behavior to the information needs, interaction styles, and media preferences of individual users in particular situations. In the context of the World-Wide Web it wouldn't make sense to delegate the task of personalizing presentations to the information providers. One reason is that information providers would have to anticipate all possible users and situations in order to provide adequate presentation formats.

In the tutorial, we will discuss recent developments in the design of intelligent multimedia interfaces that go beyond the standard canned text, predesigned graphics and prerecorded images and sounds typically found in commercial multimedia systems of today. We will show that it is possible to adapt many of the fundamental concepts developed to date in NL processing, user modeling and discourse research in such a way that they become useful for multimedia presentations as well. We will address key applications such as communication assistants for the Internet, multimedia helpware, information retrieval and analysis, authoring, training, monitoring, and decision support. There is also a peripheral aspect of personalizing user interfaces when the system personifies itself audio-visually, e.g. as an animated life-like character.

The tutorial will introduce the technology for the development of animated interface agents who play the role of communication assistants explaining, commenting and highlighting the material to be presented. The tutorial will be augmented by numerous videos and interactive demos of research prototypes and commercial applications.

Prerequisite Knowledge

The tutorial assumes no prior knowledge on multimedia interface agents, but a basic knowledge of AI concepts, including AI planning, NL processing, graphics generationand user modeling, will enhance the value of this course for participants.

About the Lecturers

Wolfgang Wahlster is the Director of the German Research Center for AI (DFKI) and a Professor of Computer Science at the University of Saarbruecken. He is a AAAI Fellow, a Trustee of IJCAII, the Chair of ECCAI and the Programme Co-Chair of ACL/EACL-97. He has published more than 150 technical papers on intelligent user interfaces. His current research includes multimodal interfaces, user modeling, and speech translation. He is a Co-Editor of the Readings in Intelligent User Interfaces.

Elisabeth Andre has been a member of the 'Intelligent User Interfaces' group at DFKI since 1989. Her research interests include: multimedia communication, intelligent user interfaces, natural-language processing and life-like characters. Since January 1994, she has been member of the executive board of the ACL Special Interest Group on Multimedia Language Processing (SIGMEDIA).

Both presenters are actively involved in numerous industrial projects dealing with various applications of intelligent user interface technology.

Last modified: Thu Feb 20 13:53:49 JST 1997