26th Annual Meeting of the Cognitive Science Society [link to conf] [Link to contact]

Tutorial Program at Cognitive Science 2004, 4 August 2004

Introduction

Topics

Conference home

Arrangements

Original call

Introduction: The Tutorials program at Cognitive Science 2004 provides participants an opportunity to gain new insights, knowledge, and skills from a range of tutorials in the field of cognitive science. Tutorial topics will be presented in a taught format and range from practical guidelines to academic issues and theory. This is the fifth year that tutorials in this format will be offered.

Tutorials will present tutorial material, that is, provide results that are established and will do so in an interactive format. They tend to involve an introduction to technical skills or methods. In this past they have included eye-tracking, statistical modelling, methods of supporting online seminars, and tools for teaching cognitive science. This year the set of tutorials is focused on a range of cognitive architectures for modeling and teaching higher-level cognition. They will include substantial review of material. The level of presentation assumes that the attendees have at least a first degree in a cognate area. All of the tutorials are related to this year's theme, Higher-order Cognition.

There is a student rate, and such students must bring their ID to show at registration.

Attendance at the tutorials does not require conference registration, but tutorial registration does not provide conference entrance. Use this form, http://acs.ist.psu.edu/cogsci2004/tutorialsonly.pdf to register just for tutorials.

Arrangements: There are five tutorials this year (one with two parts) on Wednesday 4 August (rooms to be announced on the day). They cost $65 (about 35 pounds or 55 Euros) for each half-day tutorial and $40 for students. Payment can be made using the registration site on the conference page, or can be paid for on the day (if space is available). The program includes handouts, and a tea and a coffee break (including biscuits). Lunch is available nearby in downtown Chicago. There will be a meeting of the tutorial committee and tutors after the tutorials, location to be announced at the tutorials.

Registration for tutorial attendees will be from 8.30 am on 4 August in the lobby of the River North Hotel. 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.

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

 Topics

Towards integrated cognitive architectures: Basic agent models
Bonzon, Half-day (0915-1230)
in the Westin River North, room to be announced

CHREST, a Tool for Teaching Cognitive Science
Gobet and Lane, Half-day (1400-1715)
in the Westin River North, room to be announced

ACT-R Tutorial
Taatgen, Half-day (1400-1715)
in the Westin River North, room to be announced

Bayesian models of inductive learning
Tenenbaum and Griffiths, Part 1: Half-day (0915-1230) Part 2: Half-day (2-515 pm)
in the Westin River North, room to be announced

Development of Cognitive Agents Using the COGNET Architecture and iGEN Toolset
Zachary and Szczepkowski, Half-day (0915-1230)
in the Westin River North, room to be announced

========================================================================

Towards integrated cognitive architectures: Basic agent models

Half-day tutorial (0915-1230)
in the Westin River North, room to be announced

Pierre Bonzon
University of Lausanne, Switzerland
pierre.bonzon@unil.ch

If memory is the medium of consciousness, then communication is its trigger. Therefore, if next generation models of cognition are to include essential characteristics of human intelligence such as consciouness, introspection, reflection, and the like, then they ought to include communication, and this will in time become formal models of communication. The objective of this tutorial is to expose, both in abstract and concrete terms, an approach for defining and implementing integrated cognitive architectures. This approach is based on recent results that introduce logical communication primitives within a model of non-deterministic agents with plans. A language for agent dialogues will be introduced. This language can be compiled into agent plans and executed on a virtual machine. Using a multithreaded extension of this abstract machine allows agents to engage in multiple conversations. The resulting integrated model could be used as a basis for building and experimenting with various cognitive architectures.

This tutorial is targeted at people with no prior exposure on the subject. A short overview on current agent models will be given as an introduction. The presentation will then focus on defining the various constituents of the proposed integrated model. These definitions will first take the form of procedures and/or functions, and then be implemented as Prolog programs. Although the ability to read Prolog programs will be required in order to understand the details of these implementations, no particular knowledge will be needed in order to follow the conceptual and/or theoretical developments.

Pierre Bonzon is a professor at the University of Lausanne. In the past, he has been invited professor at the University of California at Santa Cruz (1985-6) and a visiting scientist at the International Computer Science Institute in Berkeley, CA (1994-95). He is currently active as a researcher in the field of communicating agents.

========================================================================

CHREST, a Tool for Teaching Cognitive Science

Half-day tutorial (1400-1715)
in the Westin River North, room to be announced

Fernand Gobet and Peter Lane
Brunel University and University of Hertfordshire
fernand.gobet@brunel.ac.uk

CHREST (Chunk Hierarchy and REtrieval STructures) is a comprehensive computational model of human learning and perception. It has been used to successfully simulate data in a variety of domains, including: acquisition of syntactic categories, expert behaviour, concept formation, implicit learning, and the acquisition of multiple representations in physics for problem solving. The aim of this workshop is to provide participants with an introduction to CHREST, how it can be used to model various phenomena, and the knowledge to carry out their own modelling experiments.

Participants in this tutorial will:

1. Acquire a comprehensive understanding of the CHREST computational model and its relation to the chunking and template theories of cognition;

2. Explore some key learning phenomena supporting the chunking theory by taking part in a verbal-learning experiment;

3. Attempt to match the performance of a CHREST model of verbal learning with their own data; and

4. Be introduced to the implementation of CHREST in sufficient detail to begin modelling data in their own domains.

Software will be provided for the experimental settings, as well as the CHREST implementation. Supporting documents and copies of published articles will provide complete details of the theory and the model.

Please bring a laptop with you. Our plan is to have tutees work in pairs, so not to worry if you can't bring one.

Fernand Gobet is Professor of Psychology at Brunel University. He previously was Reader of Intelligent Systems at the University of Nottingham. He started working on the CHREST architecture in collaboration with Herbert Simon during his stay at Carnegie Mellon (1990-1995). In addition to developing the CHREST architecture, his scientific interests include: methodology of computational modelling; empirical research into expert behaviour, including its neural correlates; the psychology of board games; natural and artifical intelligence; and the fusion between perceptual and conceptual knowledge.

Peter Lane is a Lecturer at the Department of Computer Science at the University of Hertfordshire. He was introduced to CHREST when working as a Research Associate for Fernand Gobet at the University of Nottingham. In addition to work with CHREST, research interests include: machine learning, computational linguistics, and neural networks.

==========================================================

ACT-R Tutorial

Half-day tutorial (1400-1715)
in the Westin River North, room to be announced

Niels Taatgen
Carnegie Mellon University and University of Groningen
taatgen@cmu.edu

ACT-R is a cognitive theory and simulation system for developing cognitive models. It assumes cognition emerges through the interaction of a procedural memory of productions with a declarative memory of chunks and independent modules for external perception and actions. Since its release in 1997, ACT-R has supported the development of over 100 cognitive models, published in the literature by many different researchers. These models cover topics as diverse as driving behavior, implicit memory, learning backgammon, metaphor processing, and emotion. We have recently developed a new version, ACT-R 5.0 that is more interruptible, achieves greater across-task parameter consistency, has better mechanisms of production learning, and is more in correspondence with our knowledge of brain function. The tutorial has no prerequisite knowledge, and is intended to on the one hand give an overview of the theory, and on the other hand to offer some direct demonstration of ACT-R models. Although a half day is not sufficient to cover all material, it can wet the appetite for the full ACT-R tutorial that is available on act.psy.cmu.edu. During the tutorial, five popular research paradigms within ACT-R will be used as a vehicle to explain the architecture: Instance learning, Utility learning, Working Memory Capacity, Perceptual/Motor constrained processing, and Rule learning.

Niels Taatgen is a researcher at Carnegie Mellon University, Department of Psychology and Universitair Hoofddocent at the University of Groningen in the Netherlands. His work is focused on skill acquisition in a broad sense, including the acquisition of complex dynamic skills but also language acquisition and cognitive development. He obtained a Masters degree in Computer Science at the university of Groningen in 1989, and a Masters degree in Psychology at the same university in 1991. In 1991 he helped setting up an educational program in Cognitive Science at the University of Groningen, which is now a very successful program with approximately 50 new students each year, and which is currently ranked second within the Netherlands. Besides teaching and organization in the new program, he worked on his PhD in psychology, which he obtained in 1999 cum laude with the title "Learning without limits: from problem solving towards a unified theory of learning". In 1999 he spent a half year at Carnegie Mellon University on a NATO grant, where he and John Anderson developed production compilation, a new mechanism to learn procedural knowledge within the ACT-R cognitive architecture. In 2000, he chaired the Third International Conference of Cognitive Modeling (ICCM) in Groningen, and continued work on production compilation, applying it to learning the past tense in English, and learning the Kanfer-Ackerman Air Traffic Controller task, a complex dynamic skill. Since March 2003 he has been working part-time at both Carnegie Mellon University (3/4) and the University of Groningen (1/4). He has extensive teaching experience in Cognitive Modeling, both in regular university courses (cognitive modeling and user models in Groningen), and in short tutorials (i.e., Amsterdam Logic School, BCN Summerschool, ACT-R Summerschool).

========================================================================

Bayesian models of inductive learning

Part 1: Half-day tutorial (0915-1230)
in the Westin River North, room to be announced

Part 2: Half-day tutorial (1400-1715)
in the Westin River North, room to be announced

Joshua B. Tenenbaum
MIT
jbt@mit.edu

Tom Griffiths
Stanford and MIT

This tutorial is designed to provide an introductory morning session, and a more advanced afternoon session. You can choose to take either or both.

Many of the central problems of cognitive science are problems of induction, calling for uncertain inferences from limited data. This tutorial will introduce an approach to explaining everyday inductive leaps in terms of Bayesian statistical inference, drawing upon tools from probability theory, statistics, and artificial intelligence. We will demonstrate how this approach can be used to model natural tasks such as learning the meanings of words, inferring hidden properties of natural kinds, or discovering causal laws, where people draw on considerable prior knowledge in the form of abstract domain theories and structured systems of relations.

Bayesian models have become increasingly popular in the recent cognitive literature. This tutorial aims to prepare students to use these modeling methods intelligently: to understand how they work, what advantages they offer over alternative approaches, and what their limitations are. We also discuss how to relate the abstract computations of Bayesian models to more traditional models framed in terms of cognitive processing or neurocomputational mechanisms. The tutorial will focus on case studies of several cognitive tasks, and for each task contrast multiple models both within the Bayesian approach and across different modeling approaches.

Joshua B. Tenenbaum is an Assistant Professor in the Department of Brain and Cognitive Sciences and the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT, and holder of the Paul E. Newton Career Development Chair. Thomas L. Griffiths is currently a Ph.D. student at Stanford University, on exchange at MIT. Their research group works actively on Bayesian approaches to learning and reasoning, contributing to both cognitive science and machine learning, as well as to interdisciplinary connections between these fields. Specific interests include models of concept learning and generalization, similarity judgment, word learning, probabilistic reasoning, sequential prediction, causal induction, and the acquisition of intuitive theories. See web.mit.edu/cocosci. Together with Alan Yuille, they are currently organizing a weeklong meeting on Probabilistic Models of the Mind at the UCLA Institute for Pure and Applied Mathematics (January 2005). This tutorial has been offered before at Northwestern University.

========================================================================

Development of Cognitive Agents Using the COGNET Architecture and iGEN Toolset

Half-day tutorial (0915-1230)
in the River North, room to be announced

Wayne Zachary and Michael Szczepkowski
CHI Systems Inc.
wzachary@chisystems.com

This tutorial introduces participants to COGNET and iGEN, an integrated cognitive/behavioral modeling method and toolset used to develop cognitive agents -- software components that exhibit a level of intelligence, which mimics human thought and behavior. Cognitive agents represent the logical transition of research on human information processing into practical application. Cognitive agents also represent a new and growing paradigm for research in decision support, intelligent human-computer interfaces, intelligent tutoring, etc., which allows cognitive models to brought to bear on problems of enhancing the interaction between people and information technology in complex work environments. COGNET/iGEN integrates computational models of expert-level human cognitive processes (e.g., an emphasis on representation of recognition-primed strategies employed by experts rather than the atomic-level constructive processes employed by novices) with many practical approaches from software and systems engineering (e.g., incorporation of well-structured external software interfaces to facilitate the integration of cognitive agents into a operational software environments) to provide a theoretically based yet practical framework for cognitive agent development. This workshop introduces participants to the concepts of cognitive agents and to the COGNET/iGEN method and tools for cognitive agent development, and prepares them to undertake the development of cognitive agents applications. Attendees are invited to bring their own laptops and will be given a (MS-Windows) CD with interactive material that can be used in the tutorial.

Please bring a laptop with you. Our plan is to have tutees work in pairs, so not to worry if you can't bring one.

Dr. Wayne Zachary is founder and President of CHI Systems Incorporated, a leading cognitive engineering company headquartered near Philadelphia. He led the development of the Cognition as a Network of Tasks (COGNET) framework for cognitive task analysis of real time, multi-tasking domains, and applied COGNET to diverse domains, ranging from military command and control to patient-physician clinical encounters. He holds a Ph.D. in Anthropology and Computer Science from Temple University. Prior to founding CHI Systems, he was Associate Professor of Information Systems at Drexel University. He developed and has taught prior versions of this tutorial for the Human Factors and Ergonomics Society, the Cognitive Science Society, the Aviation Psychology Society, and the Behavioral Representation in Modeling and Simulation annual conference.

Mr. Michael Szczepkowski is a cognitive engineer at CHI Systems' Philadelphia regional office and manages the Cognitive Engineering Practice there. He holds a M.S. in Human Factors from S.U.N.Y. at Buffalo. He has used COGNET/iGEN to developed cognitive models used in numerous cognitive agents for instructional systems and decision support systems over the last decade. He developed the commercial training course curriculum for COGNET/iGEN and has instructed over one-hundred students in this course in the United States and Europe. He has also helped teach prior versions of this tutorial.

========================================================================

Co-Chairs

Frank E. Ritter (Penn State)
Frank Keller (U. of Edinburgh)

Committee members

Adele Abrahamsen (UCSD)
Fernanda Ferreira (Michigan State)
Todd Johnson (UT/Houston)
Gary Jones (Derby)
Chris Kello (George Mason)
Padraic Monaghan (Warwick)
Timothy J. Nokes (U. of Illinois/Chicago)
Bob Schewe (DDB)
Ching-Fan Sheu (Depaul)
Robert St. Amant (North Carolina State University)
Yvette Tenney (BBN Labs)
Richard Young (Hertfordshire)

Conference Overview | Local Arrangements | Registration | Program | Sponsors