Logic has been developed over the centuries as a formal way of representing assumptions about a world, i.e. it is a formalism in which problems can be formulated. It has the advantage that this formulation can be made independent of an execution mechanism or a search strategy.
Indeed, logic allows separating the algorithmic treatment of a problem from the description of the problem itself. You can reason about the knowledge expressed in logic and verify its correctness since you know its meaning.
The course aims at convincing students of this power of logic, at training them in stating a problem in logic, and at making them acquainted with logic approaches to model different kinds of knowledge (definite knowledge, disjunctive knowledge, negative knowledge, and first-order logic). In addition some forms of more advanced reasoning will be presented and situated in current research fields (possible candidates are default reasoning and abduction).
Some familiarity with algorithms and data structures. Basic knowledge of elementary set theory.
Articles and literature
Is also included in other courses
- Master of Artificial Intelligence 60 ects.
The selection of studied approaches, languages and systems can vary from year to year. For each introduced approach, language or system the conceptual foundations, the main built-in representation and problem solving features, the illustrations of the use and limitations and the indication of the key research aspects and the key applications are the important themes.
The syntax, semantics, and proof theory of the logics are studied.
The basic notions of model theory are introduced (interpretations of sets of clauses, models, and logical entailment).
The part on proof theory studies resolution, substitution, unification, top-down and bottom-up inference, and correctness and completeness of proof procedures. While studying the different logical formalisms we discuss the impact of their expressive power on the computational behaviour of the proof procedures. Also some forms of more advanced reasoning will be studied together with some relevant research aspects (possible candidates are approaches to default reasoning including closed world reasoning and abduction).
The use of the different logics for representing knowledge is illustrated by means of some applications.
Note that this course focusses on the logical formalisms and not on logic programming aspects.
Description of learning activities
This learning activity can take 2 forms depending on the number of students participating in this course. The choice concerning the learning activity has an effect on the kind of evaluation too.
Option 1: the sessions are organised as a reading club. For each session, all the students prepare the material. Every student will be responsible for guiding a part of the sessions (provide a summary and raise discussion points).
Option 2: the lecturer presents the material.
During the first lecture, all students have to be present and decisions are taken.
Each year the lecturer will provide the course material: it will be a selection of relevant chapters of a text book and/or some papers.
The selection will be announced via the home page of the lecturer: http://www.cs.kuleuven.be/~gerda .
If the course is organised as a reading club, the evaluation consists of permanent evaluation (30%) and an assignment (70%). The permanent evaluation is about the activity and creativity of the students during the reading club sessions. The assignment is excercise based. The assignment is given during the semester and has to be handed it before the start of the exams.The assignment is the same for the third examination period.
Otherwise, oral exam with written preparation. Closed book.