Verkkopedagogiikka

1
AMIS Ambient Multimodal Information Service
FP6 IP proposal Markku Turunen University of
Tampere COST 278 MC meeting, Barcelona, 10.2.2003
2
FP6/IST strategic objectives
Technology components
Integrated systems
Sectoral Applications
Source 2003 2004 Workprogramme, European
Commission
3
IPs possible implementations

• Monolithic
• partners known at outset
• tasks identified
• budget known

• Incremental participation
• tasks identified
• budget known
• not all participants are in

• Incremental funding
• objectives known
• technology roadmap known
• tasks to be completed
• budget to be changed
• participants as well

4
Multimodal interfaces

• key objectives and breakthroughs
• natural and adaptive multimodal interfaces that
  respond intelligently to speech and language,
  vision, gesture, haptics and other senses
• research tasks
• interaction between and among humans and the
  virtual and physical environment
• multilingual systems

5
Consortium 1/2

• Universities
• University of Tampere, Finland
• Bielefeld University, Germany
• Linköping University, Sweden
• IT-University of Copenhagen, Denmark
• University of Helsinki, Finland
• University of Augsburg, Germany
• University of Twente, Netherlands
• University of Brighton, United Kingdom

6
Consortium 2/2

• Research institutes and companies
• CST, Denmark
• LIMSI-CNRS, France
• Alexandra Institute, Denmark
• Knowledge S.A.
• Speech-Ware

7
Objectives

• General aim
• Produce an ambient multimodal information service
• outside the users ordinary office or home
  environment
• accessible in different physical scenarios
• adaptable to changes in the environment
• supporting information and transaction access to
  complex data
• providing natural interaction by means of
  integrated multiple input and output modalities.

8
Research activities 1/4

• impact of multimodality in ambient information
  systems
• which modalities for which tasks
• placing of I/O devices in environment
• intentional vs. unintentional modality use
• robustness requirements in ambient systems
• robustness enhancement (e.g. of speech
  recognition) via integration of modalities
• reasoning capabilities

9
Research activities 2/4

• modality integration
• methods and architectures for modality
  integration, e.g. NLP-oriented vs. machine
  learning techniques

10
Research activities 3/4

• multimodal dialogue
• dialogue modelling in ambient services
• robust processing of multimodal input embedded in
  a dialogue
• coherent communicative reactions to changes in
  the environment
• versatile multimodal generation

11
Research activities 4/4

• dynamism, modularity and portability
• storing and automatic updating of dynamic domain
  information, including a dynamic ontology based
  on emerging standards
• component parameterisation
• adaptability to user models
• adaptability to different languages and domains
• development of generic platform standards

12
More information Prof. Kari-Jouko
Räihä kjr_at_cs.uta.fi