|
My current research is focused on theoretical and applied issues of Multiagent Systems (MAS). I'm specially interested in the conflict between Autonomy and Control in agents and the use of norms to provide a solution.
Autonomy is one of the most desired properties of agents.
We want agents to be autonomous in order to be able to (proactively) take
their own decissions and to adapt to new, unexpected situations. On the other
hand we want agents to behave within some acceptable boundaries, in
order to achieve one or several goals. Therefore some control should be applied to the agents' behaviour.
Autonomy vs. Control problem:
How to ensure (control) an efficient and acceptable behaviour of a Multiagent System without diminishing the agents' autonomy?
|
This compromise between autonomy and control is present in three different levels:
- The social level: The definition of agent societies, their shared goals, the roles and accepted behaviour.
- Social Structures, Conventions, Norms and e-Institutions.
- Norm Enforcement Mechanisms.
- The agent cognitive level: The (flexible) enactment of an role by an agent.
- Cognitive Agents, Normative Agents.
- The agent physical (situated) level: only present in situated agents (robots).
- Learning and flexible behaviours during Navigation and Task execution.
My current research is therefore organized in two main lines
To test and improve the research, I'm specially interested in two application domains
My past research was focused on uses and applications of Software Agents and Multi-agent
Architectures, and was composed by two research lines:
|
Norms, e-Institutions and Normative Agents
|
|
This is my main
research line, which focuses in the concepts of norms
and institutions in order to provide normative
frameworks to restrict or guide the
behaviour of (software) agents.
The main idea is that the interactions among a group
of (software) agents are ruled by a set
of explicit norms expressed in a computational language
representation that agents can interpret. Such norms
should not completely constrain (control) the autonomy
of the agent, but guide the agent choices by defining
which behaviours and situations are acceptable (legal
) or unacceptable (illegal ).
An additional advantage of norms is that they reduce
the complexity of the environment by making the
behaviour of other agents more predictable.
A good summary
of my research objectives, approach and current research
lines can be found in the following position paper:
- J. Vázquez-Salceda.
"From Human Regulations to Regulated
Software Agents' Behaviour". (presented
in the First Round Table on Virtual Institutions and
Legal Theory, Barcelona, may 2005). [PDF
paper (english)] [PDF
slides (english)]
Recently I became
interested in the application of this line of research
in non-agent-based technologies. Currently we are applying
both institutional and organisational modelling in Service-Oriented
Architectures (GRID and Web services) as part of the
work carried out in the EU funded projects PROVENANCE,
CONTRACT and ALIVE.
Projects:
-
PROVENANCE:
This project
has defined an open provenance architecture, an architecture
for provenance systems, based on an open data model,
allowing explicit documentation of past processes
to be expressed, and a set of public interfaces allowing
the creation, recording and querying of such process
documentation.
-
CONTRACT:
The aim
of the project is develop frameworks, components and
tools which make it possible to model, build, verify
and monitor distributed electronic business systems
on the basis of dynamically generated, cross-organisational
contracts which underpin formal descriptions of the
expected behaviours of individual services and the
system as a whole.
- ALIVE:
The project
will develop new approaches to the engineering of
distributed software systems based on the adaptation
of coordination and organisation mechanisms, often
seen in human and other societies, to Service Oriented
Architectures. Such methods provide robust descriptions
of distributed systems and make the development of
complex software systems more accessible to non-specialists.
Some publications:
- J. Vázquez-Salceda.
"The role of Norms and Electronic Institutions
in Multi-Agent Systems". Whitestein Series in
Software Agent Technology, Birkhäuser Verlag
AG, Switzerland. ISBN 3-7643-7057-2 [More
info]
- J. Vázquez-Salceda.
"Normative Agents in Health Care: Uses and Challenges".
AI Communications vol. 18 num. 3, pp. 175-189. IOS
Press, 2005. [check
the Technical Report]
- J. Vázquez-Salceda,
V. Dignum, F.Dignum. "Organizing Multiagent Systems".
Journal of Autonomous Agents and Multi-Agent Systems,
vol. 11 issue 3, pp. 307-360. Springer Science, November
2005. [check the
Technical Report]
- J. Vázquez-Salceda,
F. Dignum. "Modelling
Electronic Organizations" In V. Marik, J. Muller
and M. Pechoucek (Eds.) Multi-Agent Systems and Applications
III: 3rd.
International/Central and Eastern European Conference
on Multi-Agent Systems -CEEMAS'03-.
Lecture Notes in Artificial Intelligence 2691, pp.
584-593. Springer-Verlag, Berlin, 2003.
[PDF
paper (english)]
- J. Vázquez-Salceda,
H. Aldewereld, D. Grossi, F. Dignum. "From
Human Regulations to Regulated Software Agents' Behaviour:
connecting the abstract declarative norms with the
concrete operational implementation".
Artificial Intelligence and Law vol. 16, num. 1, (March
2008). Springer. ISSN: 0924-8463. [available
online]
- H. Aldewereld,
D. Grossi, J. Vázquez-Salceda,
F. Dignum. "Designing Normative Behaviour
by the Use of Landmarks". AAMAS-05 International
Workshop on Agents, Norms and Institutions for Regulated
Multi Agent Systems -ANI@REM-, Utrecht, The Netherlands,
July 2005. [PDF
paper (english)]
- H. Aldewereld, J.
Vázquez-Salceda, F.
Dignum, J.J. Ch. Meyer. "Verifying Norm
Compliance of Protocols".
AAMAS-05 International Workshop on Agents, Norms and
Institutions for Regulated Multi Agent Systems -ANI@REM-,
Utrecht, The Netherlands, July 2005. [PDF
paper (english)]
- D. Grossi, H.
Aldewereld, J. Vázquez-Salceda, F. Dignum.
"Ontological Aspects of the Implementation of Norms
in Agent-Based Electronic Institutions" Computational
& Mathematical Organization Theory vol.
12 issue 2-3, pp. 251-275. Springer, October 2006.
[check the original Normas05
paper]
- N. Oren, S. Panagiotidi,
J. Vázquez-Salceda, S.
Modgil, M. Luck and S. Miles. "Towards
a formalisation of Electronic Contracting Environments".
In Hubner, J.F.; Matson, E.; Boissier,
O.; Dignum, V. (Eds.). Coordination, Organizations,
Institutions and Norms in Agent Systems IV. Lecture
Notes in Artificial Intelligence , Vol. 5428, pp.
156-171. Springer, 2009. ISBN: 978-3-642-00442-1 [paper
in Springer Online]
- J. Vázquez-Salceda,
S. Álvarez-Napagao "Using SOA
Provenance to Implement Norm Enforcement in e-Institutions".
In
Hubner, J.F.; Matson, E.; Boissier, O.; Dignum, V.
(Eds.). Coordination, Organizations, Institutions
and Norms in Agent Systems IV. Lecture Notes in Artificial
Intelligence , Vol. 5428, pp. 188-203. Springer, 2009.
ISBN: 978-3-642-00442-1 [paper
in Springer Online]
- J. Vázquez-Salceda,
U. Cortés, J. Padget, A. López-Navidad,
F. Caballero. "The organ allocation process: a
natural extension of the Carrel Agent Mediated Electronic
Institution". AI Communications vol. 16 num. 3,
pp. 153-165. IOS Press, 2003. [PDF
paper (english)]
- J. Vázquez-Salceda,
J.A. Padget, U. Cortés, A.
López-Navidad, F. Caballero.
"Formalizing an Electronic Institution for the
distribution of Human Tissues". Artificial Intelligence
in Medicine vol. 27 issue 3 , pp. 233-258. Elsevier,
March
2003. [PDF
paper (english)]
|
|
Learning in Multi-robot Systems
|
|
Some people define Robotics as the
research area where Artificial Intelligence meets the real world. In other
words, robots have to cope with perception (e.g., noise) and actuation (e.g.,
slippage) problems and therefore modify or even override some planned actions
(created in a higher, deliberative layer) to cope with an unexpected problem
or a safety situation.
Therefore we have an interesting instance of the autonomy vs. control
problem between the deliberative and reactive layers of a robot architecture.
In principle the deliberative layer should decide (control) the actions taken
by the robot, but the reactive layers should have autonomy enough to change
the planned actions to overcome with detected deviations or with safety problems.
This research line, coordinated by Cristina Urdiales, aims to develop a hybrid
(deliberative/reactive) architecture for autonomous behaviour in Multi-Robot
Systems composed by three levels:
- The social level, to coordinate actions between robots
and robots, and between robots and humans. This layer provides a list of
abstract goals to each robot.
- The deliberative level, where each robot creates
a plan of action, taking into account the abstract objectives from the social
level and the perceptions from the robot sensors. This layer provides then
a decomposition of partial goals (subgoals) to be achieved.
- The reactive level, which autonomously tries to achieve
each of the subgoals, solving any small problem, obstacle or deviation
that it might find and learning from its experiences to solve new siilar
problems..
A first hybrid architecture for autonomous navigation, covering the deliberative
and reactive levels, has been already implemented and tested on a Pioneer
robot with 8 Polaroid sonar sensors, where Case-Based Reasoning is used for
navigation in the reactive layer.
As an application domain, we aim to create robotic platforms to safely movematerial
and elderly and handicapped patients inside a hospital, as aprt of
the e-Tools project (see the Multiagent Systems for Medical Applications research area below).
Some publications:
- D. Isern, R.
Annicchiarico, J. Vázquez-Salceda,
C. Urdiales, C.
Caltagirone. "Applying Multi-Agent Systems and
Bio-Mimetic Navigation Architectures for Coordination
and Scheduling of Humans and Robots". 10th International
Symposium on Robotics and Applications –ISORA
2004-, Sevilla, Spain, June 2004. [PDF
paper (english)]
- C. Urdiales,
E.J. Perez, J.
Vázquez-Salceda, F.
Sandoval. "A hybrid architecture for autonomous
navigation using a CBR reactive layer". Proceedings
of the 2003 IEEE/WIC International Conference on Intelligent
Agent Technology (IAT 2003), Halifax, Canada, 13-16
October 2003, pp 225-232. IEEE Computer Society 2003.
ISBN 0-7695-1931-8. [PDF
paper (english)]
- C. Urdiales,
J. Vázquez-Salceda, E.J. Perez, M. Sànchez-Marrè
and F. Sandoval. "A CBR based pure reactive layer
for autonomous robot navigation". Proceedings
of the 7th IASTED International Conference on Artificial
Intelligence and Soft Computing, July 14-16, 2003,
Banff, Canada, pp. 99-104.
Acta Press 2003. ISBN 0-88986-367-9. [PDF
paper (english)]
|
|
Multiagent Systems for Medical Applications
|
|
This research line, leaded by Ulises
Cortés,
applies Knowledge Engineering, Machine Learning and Agent coordination mechanisms
to create agent-mediated applications in the Health Care domain.
The interest on Health Care as an application domain is because its inherent
complexity, which presents a wide variety of issues to be solved.
This line has currently two application domains:
- Assisting practitioners in the organ and tissue allocation problem (the Carrel/UCTx project)
- The use of Agent technologies to create tools for elderly and handicapped patients (the e-Tools project)
Projects:
-
Carrel/UCTx:
project in collaboration with the Hospital de la Santa Creu i Sant Pau,
in Barcelona. The project has two working lines 1) design
of an Electronic Institution (Carrel) for the exchange of human tissues
among hospitals, an Agent-mediated institution; 2) design on an Agency
(UCTx) to assist a Transplant Coordination Unit, an multi-agent system
that negotiates and communicates with Carrel.
-
e-Tools:
project in collaboration with IRCSS Fondazione Santa
Lucia (Rome) and the University of Malaga. The project
aims to design and build electronic tools that improve
the quality of life of impaired patients by increasing
their autonomy in carrying out their daily activities.
This research project created the foundations for
the EU-funded SHARE-it
project.
Some publications:
- I. Rudomín,
J. Vázquez-Salceda, J.L. Diaz de León
Santiago (eds.).
"e-Health: Application of Computing Science in
Medicine and Health Care" Research on Computing
Science vol. 5. Instituto Politécnico Nacional,
Centro de Investigación en Computación,
México 2003. ISBN 970-36-0118-9.
[Cover
(jpg)][Table
of Contents (pdf)]
- J. Vázquez-Salceda.
"Normative Agents in Health Care: Uses and Challenges".
AI Communications vol. 18 num. 3, pp. 175-189. IOS
Press, 2005. [check
the Technical Report]
- U. Cortés,
R. Annicchiarico, J. Vázquez-Salceda, C. Urdiales,
L. Cañamero, M. López, M. Sànchez-Marrè,
C. Caltagirone. "Assistive technologies for the
disabled and for the new generation of senior citizens:
the e-Tools Architecture". AI Communications vol.
16 num. 3 pp. 193-207. IOS Press, 2003. [PDF
paper (english)]
- J.
Sousa Lopes, J. Vázquez-Salceda, M. Fernández
Carmona and C. Urdiales. "Non-intrusive
sensoring and behavior analysis in residences for
the elderly". Proceedings
of the Fifth Workshop on Agents Applied in
Health Care at AAMAS'08, Estoril, Portugal, May 2008.
- J. Vázquez-Salceda,
U. Cortés, J. Padget, A. López-Navidad,
F. Caballero. "The organ allocation process: a
natural extension of the Carrel Agent Mediated Electronic
Institution". AI_Communications vol. 16 num. 3,
pp. 153-165. IOS Press, 2003. [PDF
paper (english)]
- J. Vázquez-Salceda,
J.A. Padget, U. Cortés, A.
López-Navidad, F. Caballero.
"Formalizing an Electronic Institution for the
distribution of Human Tissues". Artificial Intelligence
in Medicine vol. 27 issue 3 , pp. 233-258. Elsevier,
March
2003. [PDF
paper (english)]
- U. Cortés,
A. López-Navidad, J. Vázquez-Salceda,
D. Busquets, M. Nicolás, S. López, A.
Vázquez, F. Vázquez i F. Caballero.
“UCTx: A Multi-Agent approach to model a
Transplant Coordination Unit”. Presented
at the 3er. Congrés Català d’Intel·ligència
Artificial (CCIA 2000), Butlletí de l'ACIA
22:23-28. Vilanova i la Geltrú, October 2000.[PDF
paper compressed with ZIP (english)]
|
|
Case-Based Reasoning and Agents
|
|
This research line is coordinated
mainly by Miquel
Sànchez and focuses on the development of more powerful Case-Based
Reasoning (CBR) systems that can work in environments and topics difficult
to model with other approaches, and environments with missing information.
Another working line (the DAI-DEPUR++ project) studies the integration
of different knowledge and reasoning models (Ontologies, Rules ans Cases)
in the same system. Another uses of the CBR are also analysed, such as
using CBR to improve the reasoning and learning capabilities of Software
Agents.
Part of these work was
done in the framekork of the european funded project A-TEAM, under the
supervision of Ulises
Cortés and Miquel
Sànchez, where we developed a Case-Based Reasoning module
for environmental emergencies management.
Currently the Case-Based Reasoning module is being succesfully applied to robot navigation (see the Learning in Multi-Robot Systems line above).
Projects:
-
"Sistema Inteligente de ayuda a la toma
de decisiones para la predicción y minimización del impacto
de la actividad volcánica" ("Intelligent decission support system
for prediction and minimization of the volcano activity impact"). Project
made in collaboration with the Universidad
de las Américas de Puebla (UDLAP) and funded by CONACYT. This
project joins Geographical Information Systems' Technologies (GIS) with
Knowledge Management ones (such as CBR) to build a decission support system
to coordinate the evacuation tasks of the population arround the Popocatepetl
volcano, in México.
-
A-TEAM
(Advanced Training System for Emergency Management). European funded
project. In this project several technologies were joined to create a Distance
Learning distributed tool for environmental emergency management teams
(such as fireworkers or emergency management teams in chemical plants).
Some publications:
- J. Vázquez-Salceda, M. Sànchez-Marrè
and U. Cortés. "Using Case-Based Reasoning
to overcome high computing cost interactive simulations".
In
K.D. Ashley and D.G. Bridge (Eds.) Case-Based Reasoning
Research and Development: Fifth
International Conference on Case-Based Reasoning,
ICCBR'03, Trondheim, Norway, June 2003, Proceedings. Lecture
Notes in Artificial Intelligence 2689, pp. 581-594.
Springer-Verlag, 2003. [PDF
paper (english)]
|
|
Agent-Mediated Recommender Systems
|
|
Recommender Systems are a special
kind of multi-agent architecture based on recommending and adapting contents
or items to a given user. We have focused our work in uses of the Recommender
Ssytems to the search and adaptation of Internet contents, and the use
of Recommender Systems to improve and increase the capabilities of the
tools for Computer Support for Collaborative Work (CSCW).
This research line started
with the design and development of ACE, a Multi Agent
Recommender System for content extracted from the Web,
implemented in Java. This research was made with Alberto
Vázquez Huerga and leaded by Ramon
Sangüesa and started as a final project for
the Bachelor degree, from september 1998 to july 1999.
Since then the research goes on in
two parallel working lines: 1) adaptation of the ACE system for recommendation
in Collaborative Work environments (Colaboratorio project), where I have
been involved part-time in the specification process, and 2) design of
a generic agent-based architecture for Recommender Systems. The last one
is currenlty active and there are forthcoming publications.
Some publications:
- A. Vázquez-Huerga,
I. Barrio, J. Vázquez-Salceda, J.M. Pujol and
R. Sangüesa. "An Agent-Based Collaboratory".
Presented at the 4rt. Congrés Català
d’Intel·ligència Artificial
(CCIA 2001). Butlletí de l'ACIA 25:223-230.
Barcelona, October 2001. [PDF
paper (english)]
- R. Sangüesa,
A. Vázquez, J. Vázquez-Salceda. “Mixing
Collaborative and Cognitive Filtering in Multiagent
Systems”. Presented at the 3rd Workshop
on Agent-Based Recommender Systems (WARS 2000), Barcelona,
june 2000. [in
PDF format (english)]
|
|
