Cognitive Modeling

Lectures:
30 h

Recommended for
Students in computer science and cognitive science.

The course was last given:
Spring 00, as a undergraduate course..

Goals
To provide insight into various modeling paradigms, in particular their scope and limitations, and give hands-on experience with currently available tools and programming environments used for modeling.

Prerequisites
Introductory course in AI. Familiarity with basic AI programming techniques.

Organization
Lectures introducing basic modeling concepts and theoretical problems.
Seminars on how different modeling paradigms relate to each other.
Small modeling assignments using available modeling tools.

Contents
Cognitive neuroscience as a basis for cognitive modeling; design principles for biologically based cognitive models. Marrs three levels of analysis; should cognitive modeling be focused on resulting behavior or its underlying cognitive mechanisms? Virtues and disadvantages of connectionist versus symbolic models. Interactive activation models; processing in cascade. Unified architectures of cognition; comparison between EPIC, ACT-R and Soar. Hybrid architectures: ACT-R/PM, EPIC-Soar, COGNET.

Literature
Selected articles and book chapters.

Teachers
Rita Kovordanyi, Nils Dahlbäck, invited lecturers.

Examiner
Rita Kovordanyi.

Schedule
Fall 2000.

Examination
3 p: Active participation in the seminars and lectures. +2p: Completion and report on small
modeling assignments plus a short term paper in which an existing model is critically scrutinized. ++2p: Completion and written presentation of a small modeling project.

Credit
3 + 2 + 2 credits

Page responsible: Director of Graduate Studies