|
Introduction:The
Tutorials program at ICCM 2012 will be held on
Thursday 12 April 2012 at Franklinstrasse 28-29, at TU/Berlin.
[travel directions
to the campus]
[directions
to the 28-29 Franklinstrasse Building]
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 Conference.
Registration:
Tutorials are 10 euros per half-day for everyone (including students) and 20 Euros per day (lunch is included). You are encouraged to register
through
the
conference site. Lunch can also be purchased
separately near the tutorial site at a nice modest restaurant we have reserved (if this is unclear, we will resolve it on the day!). Attendance at
the tutorials does not require conference
registration; tutorial registration does not
provide conference entrance. You can also register
for the tutorials at the door on a space available basis [as of 5 April we have some space].
There will be a meeting of the tutorial
committee, tutors, and interested tutees after the
tutorials; location to be announced at the
tutorials. (Probably Tiergarten-Quelle or other biergarten.)
Registration for tutorial attendees will be from
8.30 am on 12 April at the ground floor of the Franklinstrasse 28-29 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. Times are S.T. (sine tempore, that is, exact times, not the sometimes used in Germany C.T., cum tempore, which means 15 min. after the posted time.)
If you have a lap top, please bring it to the
sessions, as you may work in pairs in several of the
tutorials.
The morning session includes a 15 min. coffee
break, and the afternoon session includes a 15 min.
tea break.
Once in town, use the
directions on the main conference site.
Tutorials
Developing CLARION-based Agents Using the New CLARION Library
Lynch, Full day (0900-1700)
Heuristic models of judgment and decision-making: Implementations and applications
Schooler & Neth, Half-day (afternoon: 1345-1700)
Understanding cognitive processes through language use
Tenbrink, Half-day (afternoon: 1345-1700)
Cognitive Robotics: The Symbolic and Sub-symbolic Robotics Intelligence
Control System
Kelley, Half-day (morning: 0900-1215)
Scaling models of cognition to the real world: Complexity-theoretic tools for dealing with intractability
van Rooij & Kwisthout, full day (0900-1700)
========================================================================
Developing CLARION-based Agents Using the New CLARION Library
Full-day tutorial (0900-1700)
[more
information on Clarion]
in Franklinstrasse 28-29, room to be announced
Michael Lynch, Dept. of Language, Literature and Communication, RPI
Previous tutorials on CLARION have focused mainly on presenting detailed introductions to the core theoretical concepts underlying the CLARION cognitive architecture. This time around, in addition to providing a detailed introduction to the theory, the tutorial will also focus on giving participants hands-on experience using the new implementation of CLARION -- the CLARION Library, version 6.1 (written in C#). To that end, we will introduce guidelines for setting up and using basic and intermediate aspects of the library (with detailed walk-throughs for several simulation examples) as well as present several significant new features and enhancements. As CLARION is implemented in C#, participants will learn how they can employ the CLARION library on different operating systems using either the Visual Studio or Mono development environments. By the conclusion of this tutorial, participants should be equipped with the necessary foundations to begin developing CLARION-based agents for their own applications.
Prerequisite knowledge: To get the most out of this tutorial, participants should have prior exposure to cognitive architectures (for example, having read Newell's UTC or a paper on CLARION) and have some experience using object-oriented programming languages (in particular, C# or Java). However, prior understanding of these areas does not have to be extensive because this tutorial will cover the key concepts for both CLARION and OOP.
Michael Lynch is a Professor in the Language, Literature and Communication Department at Rensselaer Polytechnic Institute, working within the Games and Simulation Arts and Sciences Program. His research includes the integration of the CLARION cognitive architecture into interactive virtual environments to facilitate the development of Intelligent Virtual Agents (IVAs). He has been directly involved in current and ongoing efforts to update both the conceptualization and implementation of CLARION and has presented tutorials on CLARION at the 32nd Annual Meeting of the Cognitive Science Society in Portland, OR and the 10th International Conference on Cognitive Modeling in Philadelphia, PA.
========================================================================
Heuristic models of judgment and decision-making: Implementations and applications
Half-day tutorial (afternoon: 1345-1700)
in Franklinstrasse 28-29 Building, room to be announced
Lael J. Schooler & Hansjoerg Neth
Max Planck Institute for Human Development
We will review simple heuristics that enable quick decisions, including take-the-best (TTB) and fast and frugal trees (FFT). TTB is a model for paired comparisons that assumes that alternatives are compared one attribute at time. FFTs classify objects through a sequence of tests. Using these heuristics, a person does not need to search for, and integrate all the relevant information to reach a decision; thus, a decision can be quickly made with little effort. Such heuristics are designed to help people make decisions in real settings, potentially achieving a high level of decision accuracy under the constraints of limited information, time, and resources.
We will illustrate these heuristics with examples from various domains to emphasize their real-world applicability. Participants will work through a series of exercises that contrast commonly used statistical techniques, such as linear regression, with their heuristic counterparts.
Prerequisite knowledge: There is no prerequisite for taking this tutorial.
Lael Schooler is a Senior Research Scientist at the Center for Adaptive Behavior and Cognition at the Max Planck Institute for Human Development. His work in computational models of memory and heuristic decision making bridges two research programs grounded in an appreciation of the adaptive value of human cognition: The program on simple heuristics explores cognitive processes that use limited information to make effective decisions and the ACT-R research program that strives for a unified theory of cognition.
========================================================================
Understanding cognitive processes through language use
Half-day tutorial (afternoon, 1345-1700)
in Franklinstrasse 28-29, room to be announced
Dr. Thora Tenbrink
University of Bremen, Germany
How can cognitive processes be accessed and understood sufficiently to enable reliable computational models? One established way of addressing internal processes is to analyze their external representations, most prominently natural language produced along with cognitively complex tasks (Ericsson & Simon, 1993). The aim of this tutorial is to familiarize young and experienced researchers with the systematic and linguistically informed analysis of language data collected in order to substantiate cognitive models. The method of Cognitive Discourse Analysis (CODA) (Tenbrink & Gralla, 2009; Tenbrink, 2010) will be introduced, which uses linguistic methods and insights to address research questions in cognitive science. One main aim is to identify particular types of linguistic patterns in the collected data that are likely to point to specific cognitive processes. The outcome of a CODA-based analysis is a validated account of systematic cognitive processes feeding directly into subsequent computational cognitive modelling. The tutorial will take the participants' current or intended projects as a starting point to address issues of motivation, data collection, CODA-based analysis, and triangulation, supplemented wherever suitable by practical exercises.
Prerequisite knowledge: There is no prerequisite for taking this tutorial.
Participants who have already collected natural language data are encouraged to bring examples as handouts or on their computers.
Thora Tenbrink is a research scientist at the Faculty of Linguistics at Bremen University (Germany), and a principal investigator in two projects in the Collaborative Research Center SFB/TR 8 Spatial Cognition (Bremen/Freiburg). Her main interest concerns the relationship between cognitive processes and linguistic representations. She is the author of "Space, Time, and the Use of Language" (Mouton, 2007), and co-editor of "Spatial Language and Dialogue" (Oxford University Press, 2009). Current research addresses wayfinding strategies, spatial communication in complex built environments, and spatial inferences derived by wayfinders from situational clues, experience, and verbal and graphical information. One further project deals with complex problem solving processes such as object assembly and analogical reasoning. See http://www.informatik.uni-bremen.de/~tenbrink/ for further information. .
========================================================================
Cognitive Robotics: The Symbolic and Sub-symbolic Robotics Intelligence
Control System
Half-day tutorial (morning: 0900-1215)
in Franklinstrasse 28-29, room to be announced
Troy Kelley
US Army Research Laboratory
This tutorial will allow participants the first hand opportunity to use the newly developed Symbolic and Sub-symbolic Robotic Intelligence Control System (SS-RICS). The system is part of an ongoing development within the Army Research Laboratory (ARL) of a robotic control architecture that was
inspired by computational cognitive architectures, primarily the Adaptive Control of Thought - Rational (ACT-R). SS-RICS combines symbolic and sub-symbolic representations of knowledge into a robotic control structure that allows robotic behaviors to be programmed in a production system format.
The architecture is organized as a goal driven, serially executing, production system at the highest symbolic level; and a multiple algorithm, parallel executing, simple collection of algorithms at the lowest
sub-symbolic level. The system should be of interest to cognitive modelers who are accustomed to developing goals and rules within ACT-R. Participants will be given a background in the theory and development of SS-RICS. Topics will include symbolic representations of knowledge, production systems, and neural networks. Additional topics will include the Simultaneous Localization and Mapping (SLAM) problem and Object Recognition problems. Tutorial participants will be given an opportunity to develop simple behavior sets in the production system using the simulated robotic environment. A background in robotics is not required.
The tutorial will be conducted as a lecture session. Additionally, students will be given the "hands-on" opportunity to develop code within the architecture. The tutorial will outline the similarities and differences between SS-RICS and ACT-R and include plenty of time for discussion. This tutorial has been given at BRIMS 2007 and BRIMS 2011. It has been modified to include more information concerning the relationship between ACT-R and SS-RICS. Additionally it has been modified to include more of the newer perceptual capabilities developed since 2005 and 2011.
Prerequisite knowledge: Some knowledge of ACT-R would be helpful but is not required.
Troy Kelley has worked at the U.S. Army's Research Laboratory's (ARL) Human Research and Engineering Directorate (HRED) since 1989. Mr Kelley's primary research interests include visual cognition, object recognition, cognitive robotics and computational models of cognition. Mr Kelley latest work has been to apply computational representations of cognition to robotics. Mr. Kelley has developed a new robotics architecture entitled the Symbolic and Sub-symbolic Robotics Intelligence Control System (SS-RICS), which is based on other cognitive architectures.
Mr. Kelley received his M.S. in cognitive psychology from George Mason University in 1998, and also received a M.S. in industrial psychology from Radford University in 1988. He has authored or co-authored over twenty-five technical publications, proceedings and journal articles. He was also the co-chair of the Behavioral Representations in Models and Simulation (BRIMS) conference from 2005-2007.
========================================================================
Scaling models of cognition to the real world: Complexity-theoretic tools for dealing with intractability
Full-day tutorial (0900-1700) [more information on the tutorial]
in Franklinstrasse 28-29, room to be announced
Iris van Rooij (co-organizer)
Department of Artificial Intelligence
Donders Institute for Brain, Cognition, and Behaviour
Radboud University Nijmegen
Johan Kwisthout (co-organizer)
Leiden Institute of Advanced Computer
Science
University of Leiden
A common theoretical obstacle encountered by computational- or rational-level models of cognition is that the cognitive capacities that they postulate appear to be computationally intractable (e.g., NP-hard or worse). Formally, this means that the computations that these models postulate consume an exponential amount of time. Informally, this means that the postulated computations do not scale in any obvious way to explain how cognitive capacities can operate in the real world outside the lab. How can cognitive scientists overcome this undesirable property of models of cognition? Over the last decade, several sophisticated complexity-theoretic techniques have been developed in theoretical computer science that can be utilized by cognitive modelers to systematically generate hypotheses about model changes or constraints that yield computational tractability without loss of the general applicability of the models. With this workshop we aim to bring these complexity-theoretic techniques to the attention of a broad audience of cognitive modelers and illustrate how they can be used to make cognitive models that scale to situations of real-world complexity.
The workshop will consist of a conceptual introduction, a number of illustrative applications, and discussion of several controversial issues. We will have hands-on exercises and aim for an interactive style of discussion. Exercises can be done using pencil and paper. More in detail, the schedule will be as follows.
Morning session (Introduction):
1. Conceptual Introduction to Tractable Modeling, by van Rooij
2. Definitions of Concepts and Techniques from Complexity Theory, by Kwisthout and van Rooij
3. Exercises for Participants to Practice the Techniques
4. Organized Discussion
Afternoon session (Applications):
5. What Does (and Doesn't) Make Deriving Analogies Hard? by Wareham
6. Does Recipient Design Make Intention Recognition Tractable? by Blokpoel
7. A Tractability Border in Natural Language Semantics, by Szymanik
8. Is Managing Multiple Goals an Intractable Balancing Act? by Reichman
For more information about this tutorial, full details of the schedule, and extra materials, please refer to their website: http://tcs.dcc.ru.nl/iccm2012/
Prerequisite knowledge: There are no prerequisites.
Iris van Rooij is an Assistant Professor in the Artificial Intelligence department at Radboud University Nijmegen and researcher at the Donders Institute for Brain, Cognition and Behaviour. Expertise: Cognitive Psychology, Formal Modeling, Computational Complexity, Philosophy of Mind
Johan Kwisthout is a lecturer at the Leiden Institute of Advanced Computer
Science at the University of Leiden. Expertise: Computational Complexity,
Bayesian Networks, Approximation, Abductive Reasoning.
Todd Wareham is an Associate Professor in the Department of Computer Science at Memorial University of Newfoundland. Expertise: Computational Complexity, Approximation, Analogy-based Processing.
Mark Blokpoel is a PhD student in the Theoretical Cognitive Science group at the Donders Institute for Brain, Cognition and Behaviour. Expertise: Communication, Action Understanding, Bayesian Modeling, Computational Complexity
Daniel Reichman is a PhD student at the Faculty of Computer Science and Applied Mathematics at the Weizmann Institute of Science. He also holds a PhD in Organizational Behavior from Tel Aviv University. Expertise: Approximation, Average-case Complexity, Decision making, Economics.
Jakub Szymanik is a post-doc at the Institute of Artificial Intelligence at the University of Groningen. He received a PhD at the Institute for Logic, Language and Computation, University of Amsterdam. Expertise: Logic, Cognitive science, Formal Semantics, Computational Complexity.
========================================================================
Important
Dates
- 1 March 2012: Camera-ready abstract, 2-page copy
due for inclusion in proceedings and
advertisements.
- 15 March 2012: Camera-ready tutorial notes
due (if we are to copy)
Committee
members
Thierry Bellet (INRETS, France)
Jim Davies (Carlton)
Maik Friedrich (TU/Chemnitz und German Aerospace Center)
Olivier Georgeon (Penn State)
Randolph M. Jones (Soar Technology)
Josef Nerb (Freiburg)
Êric Raufaste (U. of Toulouse le Mirail, France)
Nele Russwinkel (TU/Berlin)
Frank Ritter (PSU, chair)
Further contact
details:
Frank Ritter
College of Information Sciences and Technology
University Park, PA 16802
Tel: + 1 814 865 4453
General Contact:
http://www.iccm2012.com/contact-information/
last updated 10 Apr 12
|