Since the PST is located in the ISTC, where a number of well-known researchers in
environmental psycology and similar disciplines enrich us with their points of view, we have
developed particular sensitivity to the problem of designing intelligent systems able to interact
This particular aspect of human computer interaction (HCI) has taken different shapes during
the years and has stimulated us to establish solid collaborations with our collegues in a number
of research projects.
In all instances, our work has dealt with the use of planning technology in two kinds of tasks:
(a) to generate different tasks for humans (interactive intelligent solvers); (b) to solve problem
integrating human-machine abilities (mixed-initiative problem solving).
Up to now we have mostly paid attention to type-a scenarios:
in an intelligent help system that gives active support to user of a commercial software tool:
[Cesta, A., Romano, G., Planning Structures for an Intelligent Help System, in G.Salvendy,
M.J.Smith, Designing and Using Human-Computer Interfaces and Knowledge Based Systems,
pp. 767-774, Elsevier: Amsterdam, 1989];
in a multi-agent that filters e-mail messages:
[Cesta, A., D'Aloisi, D., Giannini, V., Active Interfaces for Useful Software Tools,
to appear in Y.Anzai, K.Ogawa, H.Mori (Eds.), Symbiosis of Human and Artifacts, pp.225-230,
Elsevier Science: Amsterdam, 1995 (Proceedings of the 6th Int. Conf. on Human-Computer
Interaction, Yokohama, Japan, July 9-14, 1995)];
in an interface agent that shows believable human-like behaviors
[Rizzo, P., Veloso, M., Miceli, M. e Cesta, A. (1999). Goal-based personalities and social
behaviors in believable agents. Applied Artificial Intelligence, 13, 239-271];
in a loop that help a user to high level programming of a robotic arm:
[Cesta, A., Riccucci, P., Daniele, M., Traverso, P., Giunchiglia, E., Piaggio, M., Schaerf, M.
(1999). JERRY: A System for the Automatic Generation and Execution of Plans for Robotic Devices:
The Case Study of the SPIDER Arm. Proceeedings Fifth Int. Symposium on Artificial Intelligence,
Robotics and Automation in Space (i-Sairas-99), 1-3 June 1999 (ESA SP-440) pp. 687-692].
More recently, we have given increasing attention to type-b problems:
in a first project we have investigated the problem of designing interface agents endowed
with "adjustable autonomy" features, that is the ability to release/retract responsibility
to an interface agent according to the current context of interaction:
[Cesta, A., D'Aloisi, D. (1999) Mixed-Initiative Issues in an Agent-Based Meeting Scheduler,
User-Modeling and User-Adapted Interaction, 9: 45-78];
[Cesta, A., D'Aloisi, D., Agent-Based Meeting Scheduling with Users in the Loop, Working Notes
of the IJCAI Workshop on "Adjustable Autonomy Systems", Stockolm, Sweden, August 1, pp.7-15].
In research more connected to complex problem solving we are studying how to make the user
adequaltely involved in the solution process:
in the O-Oscar architecture some active services derived from the constraint-based internal
representation allow the user to understand features of complex scheduling problems and their
[Cesta, A. Oddi and A. Susi (1999). O-OSCAR: a Flexible Object Oriented Architecture for
Schedule Management in Space Application. Proceeedings Fifth Int. Symposium on Artificial
Intelligence, Robotics and Automation in Space (i-Sairas-99), 1-3 June 1999 (ESA SP-440)
in the Mexar system we have designed quite a number of interactive services that allow
the user to gain trust in the automatically generated solutions, but also allow him to
personally refining current solutions drawing on his/her personal knowledge of the problem:
[Amedeo Cesta, Gabriella Cortellessa, Angelo Oddi, Nicola Policella, Interaction Services for
Mission Planning in MARS EXPRESS, Proceedings of the 3rd International NASA Workshop on Planning
and Scheduling for Space. Houston, Texas, October 27-29, 2002].
At present we are investigating several aspects connected to mixed initiative problem solving
like the development of measure to compare different systems, to the synthesis of explanation
for the user to have him/her involved in the constraint-based activity. A different aspect is
connected to acceptability of interlligent systems and is currently connected to our studies
In pursuing this aim several remarkable aspects have been considered: "user-centered" design
principles, interactive problem solving paradigms, evaluation metrics, usability and acceptability
studies. One of the scopes of the PST's work includes designing and testing innovative tools
and technologies to support human activities and problem solving tasks by promoting an interactive
and fruitful collaboration between human and artificial solvers.
The PST carries on this tradition by fostering multidisciplinary research whose principal
lines are in the field of human-computer interaction, user-interface software tools, cognitive
models, intelligent agents, collaborative and mixed-initiative problem solving.