:: ICCM 2006 ::

Tutorial Program
International Conference on Cognitive Modeling
5 April 2006

Early registration deadline: 1 march 2006



Conference home


Original call

Introduction:The Tutorials program at ICCM 2006 will be held on Wednesday 5 April 2006 at the CSIA (Centro Servizi Informatici di Ateneo, also known as part of H2) building (via Valerio 12, I-34127 Trieste) at the University of Trieste. [map to the campus] [map to the building from the conference center] [map of the campus showing H2] Tutorials have been held successfully at previous ICCM meetings, and are now an established feature of the conference. The format of this year's program is modelled on previously successful ICCM tutorials, and is similar to the series held at the annual Cognitive Science Society Conferences.

Registration: Tutorials cost 35 Euros (about 25 pounds or about $40) for each half-day tutorial and about $30 in Euros for students (bring ID to show at registration). You are encouraged to register through the conference site, or, if space is available, paid for on the day. Lunch can be purchased seperately near the tutorial site. Attendance at the tutorials does not require conference registration; tutorial registration does not provide conference entrance. If you are not registering for the conference, you can register for the tutorials at the door.

There will be a meeting of the tutorial committee, tutors, and interested tutees after the tutorials, location to be announced at the tutorials.

Registration for tutorial attendees will be from 8.30 am on 5 April at the ground floor of the CSIA building. It should take less than 5 minutes to get from the tutorial desk to the tutorial rooms, but please allow yourself this time to get to the room.

If you have a lap top, please bring it to the sessions, as you will work in pairs in the tutorials.

The morning session includes a 15 min. coffee break, and the afternoon session includes a 15 min. tea break.

Directions: If you fly into Trieste airport you can take a taxi or bus or rent a car to get to the city center. The bus (start: Ronchi Airport, stop: Trieste) appears to run about every half-hour, and cost under $5. Taxis are 'rather expensive', which I interpret to be about $40.

Once in town, from Trieste train station, go to University/CSIA (taxi or bus or car). Take bus line "17/" from the central train station (close to the bus station) to the University (ask the busdriver to let you off at the university, you need to buy a bus ticket before getting on the bus). Then walk to the CSIA using the maps above.

I personally recommend at taxi from the city center, which will cost about 8 euros. The tutorial site is about 5 min. up the hill from the main university steps.


agimap&emdash;A tool chain to support the modelling of the interaction level of dynamic systems
Urbas et al., Half-day (1345-1700)

Psi and MicroPsi&emdash;A novel approach for modeling emotion and cognition in a cognitive architecture
Bach, Dörner, and Vuine, Half-day (1345-1700)

Introduction to Connectionist Simulation in Social Cognition
Van Overwalle, Half-day (0915-1230)

Simulation exercises in Social Cognition with FIT
Van Overwalle, Half-day (1345-1700)


agimap&emdash;A tool chain to support the modelling of the interaction level of dynamic systems

Half-day tutorial (1345-1700) [more information]
in the CSIA, room to be announced

Leon Urbas
Technische Universität Berlin, Center of Human-Machine-Systems University
leon.urbas@ tu-berlin.de, and

Marcus Heinath, Technische Universität Berlin, Germany
Sandra Trösterer, Technische Universität Berlin, Germany
Nele Pape, Technische Universität Berlin, Germany
Jeronimo Dzaack, Technische Universität Berlin, Germany
Sandro Leuchter, Fraunhofer IITB, Karlsruhe, Germany
Jürgen Kiefer, Technische Universität Berlin, Germany

Agimap is a tool chain that supports cognitive modelling in an effective and efficient way at the interaction level (i.e., perception and execution) and is implemented for ACT-R 6.0. ACT-R models are derived from a graphical description of the interface or an intermediate high-level XML representation. The tool chain provides a graphical editor to create an XML model of the human-machine interface. XSLT is used to transform this representation to genuine model (Lisp) code. Currently, the tool chain supports code creation for ACT-R's perceptual-motor subsystem and its AGI interface. To connect the AGI interface elements with a simulated external task environment or even a real process control system the agimap-extension for cognitive architectures is used. In typical applications the amount of code generated by agimap can take up to 50% of the model's code. The approach allows efficiently dividing the physical and cognitive modelling tasks by using and connecting engineering simulation tools for the "machine" part and cognitive architectures for the cognition part what speeds up the modelling process furthermore. The development of agimap was supported by DFG and VolkswagenStiftung.

Prerequisite knowledge: We expect participants to have some general programming experience and a basic understanding of symbolic processing. Some prior knowledge of ACT-R or rule-based systems is required.

Leon Urbas is an information systems engineer by training and has a professional background in information and automation systems design for the chemical process industries. He is head of the user modelling research group MoDyS at the Center of Human Machine Systems at Technische Universität Berlin and spokesman of the DFG funded research training group prometei that investigates methods for the prospective design of humanmachine-interaction. Marcus Heinath studied media technology in Ilmenau, Germany and is member of the research training group prometei. His research interest is in integrated engineering approaches for user modelling as a design tool. Sandra Trösterer studied psychology in Graz, Austria and Braunschweig, Germany and is now research scientist at the MoDyS research group. Her research interest is in cognitively sound user modelling as a tool for hypothesis generation and usability evaluation. Nele Pape studied cognitive science in Osnabrück, Germany. She joined the MoDyS research group as research scientist in the area of first principles cognitive modelling. Jeronimo Dzaack studied computer science in Bremen, Germany. As a member of the research training group prometei he investigates on usability design and usability testing with cognitive architectures. Sandro Leuchter studied computer science in Berlin, Germany. He worked as a research scientist in the MoDyS Research Group. Since 2005 he has been a senior research scientist and group manager at the Fraunhofer Institute for Information and Data Processing in Karlsruhe. Jürgen Kiefer studied psychology in Saarbrücken, Germany. He is member of the research training group prometei and investigates modeling approaches that could explain differences in multitasking performance.


Psi and MicroPsi&emdash;A novel approach for modeling emotion and cognition in a cognitive architecture

Half-day tutorial (1400-1715)
in the CSIA, room to be announced

Joscha Bach
Institute for Cognitive Science, University of Osnabrueck

Dietrich Dörner
Institute for Theoretical Psychology, University of Bamberg

Ronnie Vuine
Humboldt University of Berlin

The Psi theory of German psychologist Dietrich Dörner sketches a cognitive architecture integrating emotion and motivation in a non-trivial way. Despite recent attempts for the inclusion of motivation into architectures, for instance Ron Sun's Clarion, most models of cognition treat emotion and motivation as additional modules that do not have a direct influence on problem solving, representation, perception and memory. The Psi theory addresses this by proposing a multithematic motivational model and a system of emotional modulators that affect all levels of cognition. Psi is a nascent architecture that has not been applied to capture a wide set of regularities, such as ACT-R, but its assumptions with respect to emotional modulation have been tested by comparing computer models with the emotional responses and behavior strategies of human subjects in a complex problem solving task.

MicroPsi is a framework that aims at implementing Psi agents on a neurosymbolic level. It features an editor and simulator for hierarchical spreading activation networks, which make up the control structures and working memory of the agents, a graphical simulation environment that allows to distribute agents over computer networks to facilitate multi agent interaction, and components to visualize the simulation world and the emotional states of the agent. The MicroPsi framework has also been applied outside our group as a robot control architecture.

The tutorial will give an overview of the theory (first half) and show the application of the framework, i.e. how to set up agents based on a hybrid neurosymbolic formalism that is supported by a graphical editor, and how to deploy multi-agent simulations in the simulation environment. We will also demonstrate the 3D viewer and the integration of agent definitions with robotic sensors and actuators to use MicroPsi as a robotic control architecture.

For further information, please visit our project homepage at www.cognitive-agents.org, where we will provide an introduction.

Prerequisite knowledge: We expect participants to have some general programming experience and a basic understanding of symbolic processing. No prior knowledge of PSI or rule-based systems required.

Joscha Bach majored in Computer Science, recieving his Diplom in 2000, since then he has worked on MAS, social simulation and robotic soccer at the AI department of the Humboldt University of Berlin. He initiated work on the MicroPsi framework in 2002. Since 2003, he has been affiliated with the Institute for Cognitive Science, University of Osnabrueck, as teacher and assistant researcher within the AI group, with focus on models of emotion and the Psi theory. He is currently finishing his PhD.

Dr. phil. Dietrich Dörner, Institute for Theoretical Psychology, University of Bamberg

1979 -1989 Professor for General Psychology at the University of Bamberg, Germany.

1989 -1991 Director of the Max Planck Project Group of Cognitive Anthropology, Berlin.

since 1991: Director of the Institute for Theoretical Psychology, University of Bamberg.

Research Areas: Psychology, Logic, Philosophy, Neurophysiology, Artificial Intelligence, Simulation of Human Activity in Complex Domains

Selected Publications

Dörner, D. (1974): Die kognitive Organisation beim Problemlösen Eine kybernetische Theorie der elementaren neuronalen Prozesse beim Denken Cognitive Organisation of Problem Solving A Cybernetical Theory of the Elementary Neuronal Processes of Thinking). Bern: Huber.

Dörner, D. (1976): Problemlösen als Informationsverarbeitung (Problem Solving as Information Processing). Stuttgart: Kohlhammer.

Dörner, D. (1983): Lohhausen: Vom Umgang mit Unbestimmtheit und Komplexität (Lohhausen: How to Cope with Uncertainty and Complexity). Bern: Huber.

Dörner, D. (1999): Bauplan für eine Seele (Blueprint for a Soul). Reinbek: Rowohlt.

Dörner, D. (2002): Die Mechanik des Seelenwagens (Mechanics of the Soul). Bern: Huber.

Ronnie Vuine is a student of Computer Science and Philosophy at Humboldt University of Berlin and former student of software engineering at Hasso-Plattner Institute, Potsdam. Since 2002 project lead and principal contributer to the software development of MicroPsi framework.


Introduction to Connectionist Simulation in Social Cognition
morning (0915-1230)

Simulation exercises in Social Cognition with FIT
afternoon (1345-1700)
in the CSIA, room to be announced

Frank Van Overwalle
Vrije Universiteit Brussel

You can register to Part I only, or Part I and II.

Part I of this workshop offers an introduction to connectionist modeling on mainstream social cognition topics such as person and group impression formation and attitude formation. Part II offers concrete hands-on experience and exercises using the FIT software on these topics. Participants need no specific knowledge of social psychology, and only a very basic knowledge on connectionist modeling is needed (e.g., what a learning algorithm is and how it works will be briefly explained/rehearsed).

Part I gives an introduction on connectionist modeling of social cognition. The tutorial will lead to a better understanding of how to model social cognition, and introduces some particular solutions to help modeling this area. Included are demonstrations of published simulation work by the author. The topics covered are listed below, but some flexibility is possible depending on the interest of the audience.

  • An introduction to the major principles of the simulations
  • Impression formation about a person, including the influence of contrast and assimilation, increased recall of inconsistent behaviors and the situational context
  • Group stereotypes and biases, including illusory correlation and accentuation
  • Attitude formation and the role of systematic versus heuristic processing

Part II of the workshop involves an introduction to the FIT simulation program (can be downloaded via the links provided), and a demonstration of the simulations that were explained in Part I using simple demos of the main principles and simulations underlying connectionist learning in social cognition. In addition, exercises will be given so that participants get some hands-on experience with these simulations. Please bring a laptop (PC only) with you and indicate this on the registration form. Our plan is to have participants work in pairs, so do not worry if you can't bring one.

Material for the tutorial:

Part I: none is required

Part II: FIT program (including introductory examples and electronic manual)

- Accompanying FIT project files

Material for preparation (if you wish) or for later reference

Part I: PowerPoint presentations

Part II: Readings and exercises

Frank Van Overwalle is a full professor affiliated with the Department of Psychology at the Vrije Universiteit Brussel. He defended his PhD in 1987 for which he received the Tobie Jonckheere Award of the Belgian Royal Academy of Sciences, Letters and Arts. His main interest is in attribution and causal learning, social cognition and connectionism (artificial neural networks). He has authored some 35 peer-refereed scientific publications, in the domain of social cognition. His recent research focuses on connectionist network models of various phenomena in the domain of social cognition at large, to demonstrate the common cognitive processes underlying many social findings. The aim is to abolish ad-hoc hypothesis building which is currently very flourishing in social psychology, and to attempt to develop a more general cognitive theory, in line with general theories of psychological information processing. This has resulted in a number of publications in top-ranking journals such as Psychological Review and Personality and Social Psychology Review. His work has received over 180 references in the combined Science, Social Science and Arts & Humanities Citation Index.


Important Dates

  • 28 February 2006: Camera-ready abstract copy due for inclusion in proceedings and advertisements.
  • 15 March 2006: Camera-ready tutorial notes due (if we are to copy)


Frank E. Ritter (Penn State)

Committee members

Nick Braisby (Open University)
Fabio Del Missier (U. of Trieste)
Glenn Gunzelman (USAF)
Lucio Inguscio (U. of Rome "La Sapienza")
Randy Jones (Soar Tech and Colby College)
Josef Krems (TU Chemnitz)
Josef Nerb (U. of Education, Freiburg)
Mike Schoelles (RPI)
Peter Wallis (U. of Sheffield)

Further contact details:

Frank E. Ritter
College of Information Sciences and Technology
University Park, PA 16802

Tel: + 1 814 865 4453 (on sabbatical at Tufts in Spring 2006, so use email please)

General Contact: iccm2006@units.it.

The Office www.theoffice.it/iccm2006

last updated 4 April 06

Conference Overview | Local Arrangements | Registration | Program | Sponsors