Definition and Technologies Knowledge Representation

Outline 1. What is Knowledge? 2. What is a Knowledge Representation? 3. Knowledge Representation in Artificial Intelligence 4. Knowledge Representation Technologies

1. What is Knowledge? data – primitive verifiable facts, of any representation information – interpreted data knowledge – relation among sets of data (information), used for further information deduction. Knowledge is (unlike data) general. Knowledge contains information about behaviour of abstract models of the world.

2. What is a Knowledge Representation? 1. A KR is a Surrogate 2. A KR is a Set of Ontological Commitments 3. A KR is a Fragmentary Theory Of Intelligent Reasoning 4. A KR is a Medium for Efficient Computation 5. A KR is a Medium of Human Expression

3. Knowledge Representation in Artificial Intelligence Logical AISearch Pattern Recognition Knowledge Representation Inference Common Sense Knowledge and Reasoning Learning from Experience Ontology Heuristics Genetic Programming

4. Knowledge Representation Technologies 1. Logic based representation – first order predicate logic, Prolog 2. Procedural representation – rules, production system 3. Network representation – semantic networks, conceptual graphs 4. Structural representation – scripts, frames, objects

4.1. Logic based Representation First Order Predicate Logic – enriched by variables, predicates, functions and quantifiers ,  – logic programming backwards-chaining implementation of inference (FOPL and resolution) question answering PROLOG: horn-clause logic, no negation, backward chaining with depth first search

4.2. Procedural Representation Production Systems – procedural representation of knowledge – in the form of if – then rules (implication as the primary representation element) – forward chaining control structure that operates iteratively – inference mechanism is firing the rules – JESS: production system implemented with Java

4.3. Network Representation Semantic Networks – particularly suited to model static world knowledge – world objects and classes of objects are modelled as graph nodes – binary relations among them are captured as edges between nodes – type of edge defines taxonomical relations between nodes, i.e. subsumption of classes and object-class membership

4.3. Network Representation Source:

4.3. Network Representation Conceptual Graphs – complete bipartite oriented graph, where each node is either a concept or a relation between two concepts – each concept has got its type and an instance – express meaning in a form that is logically precise, humanly readable, and computationally tractable

4.3. Network Representation Conceptual Graph Example Source:

4.4. Structural Representation Frames – evolution of semantic networks – a hierarchy of frames – each frame has a: a name slots: these are the properties of the entity that has the name, and they have values (a default value, a specific value, a deamon, an inherited value) – predecessor of object-oriented systems

4.4. Structural Representation Frames Example Source: ftp://ftp.cs.bham.ac.uk/pub/authors/M.Kerber/Teac hing/AI/l6.pdf ftp://ftp.cs.bham.ac.uk/pub/authors/M.Kerber/Teac hing/AI/l6.pdf

4.4. Structural Representation Scripts – description of a class of events in terms of contexts, participants, and sub-events – knowledge base representation in terms of the situations that the system is supposed to understand – restaurant script