"Cotutelle de thèse entre
École de Technologie Supérieure et l'université de Versailles"
of a Web Service Component for Interaction Fusion
interaction, fusion, scenario, events, ontology, Web service.
Our laboratories (MMS/ETS and LISV/UVSQ) work on the design of
architecture for interaction. These interactions use sensors network of
modalities in several contexts. For example, to interact with a computer in a
noisy environment (context), it is better to use a keyboard or a touch screen
instead of using vocal commands where misunderstanding will occur. To design a
component of the architecture, Researchers work on different technologies as
Multi Agent Systems from the AI domain or webservices  from telecommunication
domain, both of them have the capability of acting and communicating
A recent domain called ambient intelligence is composed of a
ubiquitous network with awareness able to interact with a user. This network is
connected to all sensors and actuators; it can discover and manage them.
Standards  exist and must be respected for better interoperability, lower
cost and Sustainability.
The aim of intelligent component design is to realize:
of sensors information under uncertainties, accuracy,
temporal spaces getting primary events to build complex
events (composed events named "scenario"),
detection of context and awareness,
theories and models.
Machine learning & IA techniques can be used as fuzzy rules,
Bayesian networks, knowledge representation and pattern matching is helpful.
Many of applications are possible and integrated in
research projects: Home assistant, robotics, intelligent network, security
alarms, supervision, games ...etc.
Candidate objective is to specify and develop a fusion component
for interaction in a network of multimodal webservices. This component must
create and refine scenarios receiving multimodal events in the EMMA W3C Standard
. XML specification for EMMA, an
Multimodal Annotation Markup
language, is responding to the documented requirements. This markup language is
intended for use by systems that provide semantic interpretations for a variety
of inputs, including but not necessarily limited to, speech, natural language
text, GUI and ink input. In addition, modality specific XML languages can be
studied such as VoiceXML  for voice recognition, MPEG-7 for Video annotation,
and so on. Primary events and composed events (scenarios) will be stored in a
semantic ontology  using specific rules. All combinations of primary events
will be defined as models to help the composition. The result of the process
will be a situational/contextual meaning.
This thesis will be done in 4 parts:
state of the art
define a domain ontology, combination models and rules for
use of W3C standards
To prove the quality of the study, candidate
will develop a Home "Assistant" application to help old people or people with
disabilities. One of the solutions can be to recognize user’s activities, infer
a situation and execute the most relevant action , . For example, abnormal
situation is detected if the person is not moving and not eating (did not use
the stove), screaming because he needs help, not answering any of the system
questions, and so on. Normal situations are stored when the user usually repeats
the same actions.
 SmartWeb Handheld - Multimodal Interaction with Ontological
knowledge bases and Semantic web services - Daniel Sonntag - 2006 (can be
downloaded on Google)
Other related publications:
Candidate must have computer science research
must be a plus.
Pr. Chakib Tadj (École de
Technologie Supérieure, Quebec University):
Pr. Amar Ramdane Cherif (Versailles University, France):