SluitenHelpPrint
Switch to English
Cursus: NWI-I00054
NWI-I00054
Cognition and Representation
Cursus informatieRooster
CursusNWI-I00054
Studiepunten (ECTS)6
CategorieMA (Master)
VoertaalEngels
Aangeboden doorRadboud Universiteit; Faculteit der Natuurwetenschappen, Wiskunde en Informatica; Informatica en Informatiekunde;
Docenten
Coördinator
dr. J.J. Sarbo
Overige cursussen docent
Docent
dr. J.J. Sarbo
Overige cursussen docent
Contactpersoon van de cursus
dr. J.J. Sarbo
Overige cursussen docent
Collegejaar2016
Periode
KW1-KW2  (29-08-2016 t/m 29-01-2017)
Aanvangsblok
KW1
Onderwijsvorm
voltijd
Opmerking-
Inschrijven via OSIRISJa
Inschrijven voor bijvakkersJa
VoorinschrijvingNee
WachtlijstNee
Plaatsingsprocedure-
Cursusdoelen
  • Theory:
    Making acquaintance with a theory of signs.
    A definition of a model of cognitive activity.
    An introduction of 'naive' logic.
  • Application
    An analysis of case studies in various knowledge domains.
  • Evaluation
    A comparison of the properties of formal and human interpretation.
Inhoud

The promise of Artificial Intelligence research that the computer will become, in matters of information processing, an equal partner of man has not been realized so far. The bottleneck is the used knowledge representation that cannot deal with the adroitness of the human mind.

In this course we study a model for knowledge representation that is cognitively based, grounded in a theory of sign, and applicable to human and artificial information processing alike. By virtue of the fundamental character of cognition and interpretation, this model can be applied uniformly to knowledge modeling in different domains. Experimental evidence from neuro-physiological research shows that human information processing may work according to the same principle.

Modeling is a subject of study in many courses in Computer Science. This course generalizes modeling in a process of conceptualization. A characteristic property of this process is the use of a subset of predicate calculus, called `naive' logic, which can be shown to be present in natural language processing, reasoning, and inductive theorem proving (by humans). It can be essential for designing efficient human-computer interfaces as well.

Conceptualization as a process is assumed by traditional modeling as well, but a specification of the events of this process is usually omitted. This course provides a model of that process (and its events) and reveals the potential of this common element of human information processing, in different domains of knowledge. The process model developed suits a computational implementation also in combination with machine learning.

Bijzonderheden
• Weekly assignments distributed via Blackboard. Solutions submitted via drop-box on Blackboard. Deadline: 9a.m. on the day of the tutorial class (werkcollege).
• Information over grades: see Blackboard.
Onderwerpen
• Sign interpretation
• Conceptualization as a process
• Naive logic
• Syntactic language modeling
• Reasoning and mathematical modeling

• Knowledge summarization
Toetsinformatie
The theory of the course is tested in two parts, in a mid-semester and an end-of-semester test. The grades for both tests must be at least 5.0. The final grade is defined by the average of the two tests (which can be 5.0 hence unsatisfactory).
This value can be increased by 0.5 if the average grade for the assignments is 7.5 or higher.
Voorkennis
Basic knowledge in propositional logic, as well as in deductive and inductive reasoning (including mathematical induction) is required.
Literatuur
J. Sarbo, J. Farkas, A. v. Breemen, Knowledge in Formation: A Computational Theory of Interpretation, ISBN:978-3-642-17088-1, Springer
Werkvormen

• 32 hours lecture
• 32 hours problem session
• 104 hours individual study period
Extra information teaching methods: This course makes use of problem directed education. Assignments involve some research and theory formation by the students. For each assignment a solution is developed in an interactive fashion in class. Assignments are representative for the tests.
Verplicht materiaal
Boek
J. Sarbo, J. Farkas, A. v. Breemen, Knowledge in Formation: A Computational Theory of Interpretation, ISBN:978-3-642-17088-1, Springer
Werkvormen
Cursusgebeurtenis

Hoorcollege

Werkcollege

Zelfstudie

Toetsen
Tentamen
Weging1
GelegenhedenBlok KW2, Blok KW3

SluitenHelpPrint
Switch to English