Welcome to the COGENT tutorial. This tutorial aims to teach the use of
COGENT by developing three "research programmes". A COGENT research programme
is just a set of related cognitive models. The models in the first research
programme are relatively basic. They relate to a very simple-minded view of
categorisation, and are not intended to be cognitively plausible. The models
in the second research programme perform multi-column addition. These models
are more complex than those of the initial research programme, and illustrate
several key aspects of the COGENT approach. Their psychological plausibility,
though greater than the initial models, is still, however,slight. The models
in the thrid research programme are more advanced. These models concern
behaviour in a medical diagnosis task for which empirical data exists, and
several of the models have been published (see Fox & Cooper, 1997; Cooper & Fox, 1997; Yule, Cooper & Fox, 1998).
The structure of the tutorial is as follows:
This tutorial can be found on the web at:
This overview of the tutorial.
Background information on what COGENT is, and what COGENT isn't.
- Research Programmes
A brief outline of the concept of a research programme as employed by COGENT,
along with a description of the functionality provided by COGENT for
dealing with research programmes.
- The Graphical Interface
A description of how to use the graphical front-end provided by
COGENT to draw a box/arrow diagram. An example toy model (of
exemplar-based categorisation) is used to illustrate the concepts.
- Putting Flesh on the Bones
The toy model of categorisation is fleshed out to give a complete COGENT
- Fundamentals of Processing
An example led description of the processing cycle. The toy model of
categorisation is run to illustrate the interaction of boxes within
- More Flesh
Some improvements to the toy model of categorisation are considered.
A brief tour of the representation language used by COGENT.
- Pattern Matching
An introduction to unification, the underlying pattern matching mechanism
that is used by COGENT to guide processing.
- Notes on Methodology
Some information about theoretical and methodological
commitments embodied in COGENT.
- A Model of Multi-Column Addition
A detailed description of a simple model of multi-column addition, illustrating the use of object
properties within COGENT.
- Model 1: A Propositional Model of Medical Diagnosis
A step by step guide to the construction of a model of the performance of
a task for which real psychological data is available.
- Model 1a: Elaborating the Experimental Environment
An elaboration of Model 1 in which COGENT is used to model the
experimental environment in which the subject is placed, and in
which various COGENT facilities are used to collect and collate
- Model 2: A First-Order Model of Medical Diagnosis
A generalisation of Model 1 obtained by a representational change.
Model 1 uses propositional representations of the subject's knowledge.
Model 2 uses first-order representations, thus significantly reducing
the apparent complexity of the model.