The Knowledge Representation Framework (KR Framework) provides the basis for a number of research activities in the CAISOR agenda. The present page describes the KR Framework for a reader who has a knowledge in Computer Science and Artificial Intelligence. A separate menue item provides a lay person's introduction.
The KR Framework defines various aspects of a representation for structured information, ranging from simple facts to complex formulas and scripts. This representation is used both as a publication language in KRF-based lecture notes and technical reports, and for markup and scripting languages in KRF-based software.
In order to see the point with the KR Framework it is therefore best to see how it is used in tutorial texts and in software documentation, and not merely to look at the formal definition of the framework. Further down on this page are links to lecture notes that have been used for a course on "Artificial Intelligence and Lisp" that is given for students in their third and fourth year of programs towards a bachelor's degree in engineering at Linköping University. Although these notes have primarily been written for use in the course, they will also be a way of showing the character of the KR Framework. The lecture note that is listed as PM-krf-009 is the natural introduction.
Two advantages stand out in a comparison between this way of teaching the subject, and the ways that it is usually taught. First, the use throughout the course of a single notation that is strictly defined, easily readable and sufficiently expressive, makes it possible to cover much more ground than what is otherwise the case. Secondly, the character of the notation is such that it can be used not only for the theoretical parts of the topic, but also for writing out examples of somewhat nontrivial size, and for use in the computer exercises that accompany the course as shown on the course webpage.
Besides these course materials, there is also software documentation for the Leonardo system, which is an implementation of the KR Framework. It is listed in the separate list of Leonardo documentation. Likewise, our recent research is listed in a separate list of recently published KR research articles and, for research on other topics, under their respective tabs in the hyperpage for the CAISOR agenda.
A few of the lecture notes and reports listed below are incomplete, and all are subject to revision from time to time in order to keep up with the developments. One obvious incompleteness in most of the lecture notes is that they presently lack references to the original sources for the methods and results that are described here. I hope to be able to correct this shortcoming fairly soon.
| PM-krf-009 | Knowledge Representation Framework: Overview of Languages and Mechanisms |
| PM-krf-011 | Compendium of Programming Techniques for Knowledge-based Autonomous Systems. Part I: List Processing |
| PM-krf-016 | Compendium of Programming Techniques for Knowledge-based Autonomous Systems. Part II: Managing Information Aggregates |
| PM-krf-012 | Compendium of Programming Techniques for Knowledge-based Autonomous Systems. Part III: Intelligent Autonomous Agents |
It is intended to add lecture notes on Lisp and functional programming, including e.g. partial evaluation. The Part II lecture note is incomplete.
| PM-krf-007 | An Introduction to Decision Trees and Probabilistic Causal Networks |
| PM-krf-015 | Reasoning about Actions and Action Planning |
| PM-krf-017 | Reasoning About Actions II |
| PM-krf-018 | Description Logic and Defeasible Inheritance |
It is intended to add similar lecture notes for several additional topics.
| PM-krf-014 | The Goals of Artificial Intelligence Research. A Brief Introduction |
| PM-krf-020 | Resume of a Course on Artificial Intelligence and Lisp |
These are intended to be read at the beginning and at the end of the first course, respectively.
The following note is intended to be read before the 'First Course' lecture notes for those course participants that do not already know the aspects of logic that are covered in the note. It therefore covers topics that are not properly speaking part of AI or KR, but which are needed as background. In doing so it uses KRF notation which facilitates its use in a course that is based on the lecture notes shown above.
| PM-krf-013 | Notes on Propositional and Predicate Logic for the Knowledge Representation Framework |
At some point we should add a similar lecture note on elementary probability theory, for use as background e.g. for the lecture note on decision trees.
These notes are intended to be read after the 'First Course' material in order to provide additional depth. Unlike the case for the First Course, there is no ambition to provide systematic coverage of an area here; we will simply add lecture notes to this list when there has been occasion to prepare one. (Only one note available at present).
| PM-krf-005 | A Simple Explanation of Circumscription |