Elisabeth André#

Short laudatio by Reinhard Wilhelm#


Prof. André belongs to the key researchers in the area of multimodal interfaces and her work is technically highly advanced. She succeeded in publishing novel work in well-recognized journals, books and conference proceedings and has been serving on a striking number of editorial boards. The fact that she was the Program Chair of prestigious conferences, such as International Conference on Intelligent User Interfaces, shows how much she is respected as a leader in the field.

Most influential was Prof. Andrés work in the area of embodied conversational agents. Such agents play an increasingly important role in computer games, educational software, electronic commerce and many other applications. Already in 1997, Prof. André organized one of the first international workshops on embodied conversational agents in conjunction with IJCAI in Tokyo, Japan. The workshop resulted into a Special Double Issue on Animated Interface Agents of the Applied Artificial Intelligence Journal, the first comprehensive collection of work in this area. Her own significant contribution in this area included the development of a plan-based approach to automatically generate multimodal dialogues for virtual agents (IUI 2000 – Best Paper Award).

While the integration of gestures, mimics and speech is still mostly based on the intuitions of the system developers, Prof. Andrés work is characterized by a thorough empirical foundation. To ground her work in empirical data, she spent a significant amount of effort on the collection of corpora capturing human behaviors as a basis for the modeling of virtual agents’ behavior. In collaboration with Japanese partners (Prof. Yukiko Nakano from Seikei University and Prof. Toyoaki Nishida, Kyoto University), Prof. André conducted one of the most extensive corpus studies in the field to generate culture-specific behaviours in agents. This work formed the basis for the simulation of culture-specific behaviors in agents. Culture is not only reflected by the agents’ visual appearance, but also by their behaviors, such as mimics, gestures, speech and body posture. Nevertheless, there are hardly any research projects that explicitly model culture-specific behaviors. One of the reasons is the fact that the acquisition of knowledge necessary for behavior modeling requires an enormous amount of effort. To simulate culture-specific behaviors, Prof. André succeeded in mastering two great challenges: (1) the development of a parameterized approach to plan culture-specific behaviors for virtual agents on the basis of Bayesian Networks (2) the development of intuitive and natural forms of interaction to communicate with culture-specific agents in a simulation environment.

Research on virtual humans must draw heavily on psychology and communication theory to appropriately convey verbal and nonverbal behaviors. In this context, I would like to mention that Prof. André has undertaken a number of very interesting interdisciplinary collaborations to develop educational agents for children. A prominent example is FearNot!, a novel computer game, which was designed for school pupils in order to increase their awareness of bullying and prevent bullying in schools. Most noteworthy is the fact that FearNot! was not only evaluated in a laboratory-like settings, but also successfully tested with nearly 1000 school pupils both in Germany and the UK.

More recent work by Prof. André includes the development of more robust input recognition methods. Significant achievements in this field include:
  • the simulation of gaze behaviors for virtual agents on the basis of studies of human-human interaction (IVA 2009 paper – Best Paper Finalist)
  • the development of intuitive interfaces to generate mimics for virtual characters using game pads and data gloves (GALA 2008 – Jury Award)
  • the development of various toolboxes for emotion recognition: a) AuBT Biosignal Toolbox AuBT, which provides techniques to analyze physiological signals, b) EmoVoice, a framework for real-time vocal emotion recognition, c) Smart Sensor Integration SSI a toolbox for the synchronized recording and analysis of multiple human-generated multimodal signals, including eye gaze, speech, physiological data, gestures

The impact of Prof. Andrés work is also demonstrated by the fact that the software developed in her lab is being used by many research institutes world-wide, among other things in large international projects, such as EU FP6 CALLAS, EU FP7 Metabo and EU FP7 Iris.

The impressing number prizes her students won at international conferences show that she is able to inspire students and increase their motivation.

Any further pages in alphabetic order of their title as created by you.
#

Just click at "Create new page", then type a short title and click OK, then add information on the empty page presented to you (including maybe a picture from your harddisk or a pdf-file by using the "Upload" Button) and finally click at "Save".
...no Data available yet!

Imprint Privacy policy « This page (revision-3) was last changed on Wednesday, 10. November 2010, 16:00 by Kaiser Dana
  • operated by