Intelligent Integration and Use of Text, Image,
Video and Audio Corpora
It is now possible for sizeable corpora of Text, Image, Video and
Audio (TIVA) to be made available for research and applications. In
contrast with text processing, however, few effective methods exist for
understanding, or even searching, the content represented by images or
video and audio recordings. Intelligent, content-understanding
programs can greatly improve the usefulness of these huge quantities
of existing material. By collecting and intelligently integrating
several of these media sources, opportunities are opened up for novel
applications of existing techniques in AI, and for the development of
new AI technologies. This symposium is intended to explore current AI
research and potential future contributions of AI in making use of
TIVA sources.
The symposium is intended to foster:
* Exploration of AI techniques for creating intelligent tools for the
collection, cataloging and exploration of mixed TIVA materials.
* Discussion and presentation of experimental results concerning the
simultaneous use of information from text, image, video and audio
channels. How can integration of these sources support intelligent
behaviour and machine learning in ways that would not be possible
using any single medium alone?
In particular we want to encourage the exchange of research that
crosses the boundaries between vision, speech processing, language
processing, knowledge representation, intelligent agent design and
information retrieval, allowing participants to share ideas and effective
approaches to the problem of making TIVA sources more accessible and
useful both to people and to other intelligent programs.
For more information see: http://www.informedia.cs.cmu.edu/sssaaai/
Submission Information:
Potential participants should submit a short but complete paper
(two to seven pages) describing approaches to solving problems
in the intelligent integration of TIVA sources. We are particularly
interested in papers describing how integrated TIVA sources are
being used in functioning systems.
We encourage the presentation, for discussion, of work-in-progress,
especially where such work is supported by empirical or theoretical
evidence. Postscript or hardcopy submissions should be submitted to
the following address. We may require submission of html versions of
accepted papers. The submission deadline is Oct 25th 1996.
Michael Witbrock or Alex Hauptmann
Carnegie Mellon University,
School of Computer Science
5000 Forbes Ave Pittsburgh, PA, 15213
Email: witbrock@cs.cmu.edu, alex@cs.cmu.edu
Phone: 412 268 6247 or 412 268 1448
Fax: 412 268 5576
Organizing Committee
Alex Hauptmann (co-chair), Carnegie Mellon (alex@cs.cmu.edu),
Marti Hearst, Xerox PARC (hearst@parc.xerox.com),
Gareth Jones, Cambridge University (gjfj@eng.cam.ac.uk),
Mark Maybury, The Mitre Organization (maybury@mail01.Mitre.org),
Behzad Shahraray, AT&T Labs - Research (behzad@dnrc.bell-labs.com),
Michael Witbrock (co-chair), Carnegie Mellon (witbrock@cs.cmu.edu)