Knowledge Organization and Representation under the AI Lens
Jian Qin
Table 1 Similarities in goals, methods, and functions between KO and KR paradigms.
Paradigm Goals Methods Functions
KO Integration Organize the knowledge universe Categorize, classify generalize, synthesize Represent knowledge in publications and organize knowledge about nature and/or society
Disintegration Organize the knowledge in a domain Categorize, classify, generalize, synthesize, model Represent knowledge in data and publications in a domain
KR Production rules Represent knowledge in condition-action pairs to solve problems Use forward chaining algorithms to execute condition-action pairs Represent fragmentary knowledge in entity-attribute-value (triple) format
Semantic networks and frames Represent semantic relations between concepts Express semantic relations in triples Connect knowledge nodes through attributes or slots to form a knowledge graph
First-order logic Formalize qualifier construction in natural language Express declarative propositions using the first-order logic syntax and semantics Produce a set of axioms for reasoning