Virtual communities on the Internet
The information society can be characterized by a
marked change in businesses and markets: the transformation of business
models is due to the emergence of the Internet, in general, and most
notably, to the the new players that are the "virtual communities". The
most clearcut cases concern the music industry with its peer-to-peer
exchanges (Napster, Kazaa, etc.) that allow web users to share music
files on the Internet, the software industry with the development of
freeware, the media industry with communities of web users sharing their
opinions on a range of issues, or games communities or meeting rooms in
which virtually incarnated people socialize and swap or sell digital
goods and services.
Nicolas Auray,
David Bounie, Michel Gensollen and
Marc Bourreau are studying the ways these virtual communities alter
market and business practices by organizing demand on the basis of a
liberal and free exchange of information and knowledge. They are
endeavouring to develop a common body of knowledge and terminology and
to define the virtual public spheres in which people voice their
opinions.
Modeling spontaneous language
In order to predict new user practices and offer new
services, it is essential to better Research at ENST understanding
human communication. When they recount something, individuals calculate
the value of what they say in terms of probability and impact. When
people develop an argument, they test out and restore a logical
coherence between their beliefs and their desires. The purpose of the
work of Jean-Louis
Dessalles is to simulate such procedures and create
a model for them.
Another fundamental problem consists in
understanding how egoistic agents can manage to communicate honestly.
Multi-agent simulations reveal that honest communication is stable when
it is used to display informational dimension of the individuals
concerned.
Automatic processing of natural languages
The Internet explosion has led to a considerable
increase in the sources and volumes of textual information available
online, underlining the need for automatic processing tools to handle
this information.
Research conducted by François Yvon in
this
area aims to improve the tools currently available, especially through
the use of automatic learning technologies data-based: morphosyntactic
labeling, semantic analysis. Such technologies ultimately contribute to
the final user applications: voice applications, information search and
extraction, automatic translation, etc.
Towards a semantic web
The Web of the 21st century will be "semantic" and
socially structured or it will be drowned in noise. Today, the only
effective web search seems to be based solely on linguistic
engineering, as was the case when people were just starting to look for
information in major text corpuses. However, entrusting one’s
production/publication of resources for the Web to such search engines
alone is tantamount to giving up, if the additional effort is not made
to annotate these resources by using controlled or indexed terms in
widely accepted information
structures (thesauri, meta databases and other ontologies). Such
efforts will be generously rewarded by the capabilites of new semantic
search engines relying on knowledge engineering and by developing
networks of recognition and trust between peers.
Research into such networks has been undertaken by
Jean-Marc Saglio in
partnership with France Telecom R&D.