SAC
2006 Keynote Speakers
|
Intelligent
agent technology: Human-machine natural
dialogue and inter-agent communication
Dr.
David Sadek
France Télécom R&D
david.sadek@francetelecom.com
|
 |
Date:
Monday April 24, 2006
Abstract
The intelligent agent paradigm has generated
such a tremendous interest in today's world
of R&D, and so vast is its range of potential
applications, that it is already seen as the
source of a new technological revolution that
will encompass application domains from access
to information, to games, including resource
monitoring, personal and public digital assistants,
natural language interfaces, intelligent intermediation,
e-business, training and education, and so
on. This interest is particularly emphasized
by the inter-connection of networks and the
inter-operability of software and services.
The agent approach aims to introduce the required
intelligence into information processing.
The idea is to produce a changeover from software
that provides functions to software that offers
services. The latter will be more user-friendly,
simpler, richer and more easily adaptable,
precisely because more intelligent and mastering
the semantics of the information and the functions
being handled. The semantics technologies
enable to implement services in which the
system, namely the software agent, is capable
of carrying on a natural dialogue with the
user. Associated with a permanent multimedia
connectivity, such technologies will, in the
coming years, offer to the user a coherent
view of a seamless inter-service and inter-media
continuum. As soon as the market begins to
witness generic agent technologies along with
standards for ontologies and knowledge representation,
this will mark the real technological leap
forward introducing a deep change of telecommunication
use and market. The present scientific, technical
and industrial landscape holds every indication
that these fundamental changes will take place
progressively over the very next years.
This talk particularly focuses on why and
how the concept of cognitive agent provides
a relevant approach to model and to implement
intelligent human-machine natural dialogue
systems. The rational dialogue agent model
comes within this approach, which also allows
for handling interactions between software
agents according to the same principles. This
model is at the basis of the Artimis technology
and also at the origin of the FIPA ACL standard
communication language. Several operational
Artimis based applications have been developed
and commercial large scale services have begun
to be deployed. Beyond its promising potential,
the presented approach opens new research
perspectives in the domains of the design
of artificial intelligent behaviors and the
organization of agent societies. Several demonstrations
illustrating the different issues tackled
in this talk will be given.
Speaker's Bio
David Sadek is chair of the research program
on intelligent agents and natural interactions,
at the R&D Division of France Telecom.
He is also head of the industrial leverage
unit for advanced human-computer dialogue
based services. His work on formal models
of cognitive agents and natural communication
is at the basis of the first world-wide effective
generic technology of rational agents and
dialogue systems, which he has created with
his team. This leading edge technology, which
has entered the industrial process for large-scale
commercial services, is already deployed in
several applications of internet virtual assistants
on France Telecom portals. Also, his work
on the semantics of communicative actions
gave birth to the FIPA ACL standard for an
inter-agent communication language. In 1999,
he received, together with his team, the France
Telecom Innovation Prize for the Artimis rational
dialogue agent technology. In 2002, he was
awarded Medal Blondel for his contribution
to Cognitive Science and Artificial Intelligence.
==========================================================================
Date:
Wednesday April 26, 2006
Abstract
One of the fundamental challenges in program development,
from large applications to embedded code, is to
be able to develop, in the shortest time possible,
programs that are efficient and correct. We argue
that in order to achieve this goal, in addition
to factors such as better programming languages,
substantially improved functionality is needed from
programming environments. We present a novel program
development framework which uses "Abstract
Interpretation" as a fundamental component.
Abstract interpretation is a technique which has
allowed the development of very sophisticated program
analyzers and transformers, which are at the same
time provably correct and highly practical. The
framework that we present uses this type of program
analysis to obtain information about the program.
This information is then used to validate programs
with respect to partial specifications written using
assertions, to detect and locate bugs, to simplify
run-time tests, to perform high-level program transformations
(such as specialization, parallelization, or resource
usage control), and to enrich mobile code with safety
certificates. The system can reason with much richer
information than, for example, traditional type
declarations. This includes pointer aliasing, shapes,
exceptions, determinacy, abounds on resource consumption
(such as computational cost or sizes of data in
the program), etc. CiaoPP, the preprocessor of the
Ciao multi-paradigm programming system is an implementation
of this framework and will be used to illustrate
these ideas.
Speaker's Bio
Manuel Hermenegildo received his Ph.D.
degree in Computer Engineering from the University
of Texas at Austin, USA. He is currently Full
Professor of Computer Science at the Technical
University of Madrid (UPM) and also holds
the P. of Asturias endowed chair in Information
Science and Technology at the U. of New Mexico,
USA. Previously he was project leader at the
Microelectronics and Computer Technology Corporation
(MCC) research center and Adjunct Professor
at the CS Department of the University of
Texas at Austin.
He has published over 150 refereed papers
in the areas of programming languages, program
development tools, advanced compilation techniques,
abstract interpretation, parallelizing compilers,
parallel and distributed processing, and artificial
intelligence. He has given more than 20 invited
talks and tutorials in major conferences on
these topics. He is also area editor of the
Journal of Applied Logic, editorial adviser
of the Journal of New Generation Computing,
and has been area editor of ACM Transactions
of Programming Languages and Systems and of
Theory and Practice of Logic Programming.
He is also the elected president of the International
Association for Logic Programming, and a member
of the executive committee of the European
Association for Programming Languages and
Systems, as well as of several other international
committees, and serves as national expert
at the OECD. He is also a member of the European
Union's high-level advisory group in information
technology (ISTAG).
|
