KI-Glossar
Das vollständige Wörterbuch der Künstlichen Intelligenz
Knowledge Graph RAG
Augmented retrieval architecture that integrates structured knowledge graphs to enrich context and improve the accuracy of responses generated by language models.
Triple RDF
Fundamental data structure of the semantic web composed of a subject, predicate, and object, enabling the representation of relationships between entities in a knowledge graph.
Ontology
Formal and explicit specification of a shared conceptualization that defines the types, properties, and relationships between entities in a particular knowledge domain.
Entity Resolution
Process of identifying and merging matching entities from heterogeneous data sources to create a unified and consistent view in knowledge graphs.
Graph Embedding
Vector representation technique for nodes and edges of a graph in a low-dimensional continuous space, preserving structural and semantic relationships for ML computations.
SPARQL
Standardized query language for RDF databases, allowing extraction and manipulation of data stored in knowledge graphs with a syntax similar to SQL.
Cypher Query Language
Declarative query language specifically designed for property graph databases, using ASCII syntax to visually represent graph patterns.
Knowledge Graph Completion
Task of predicting missing relationships and absent entities in an incomplete knowledge graph, using inference and machine learning techniques.
Vector-Graph Hybrid Retrieval
Approach combining vector semantic search and structured graph search to improve the relevance and accuracy of retrieved documents in RAG systems.
Named Entity Recognition
NLP process of identifying and classifying named entities (people, places, organizations) in unstructured text, a crucial step for automatic knowledge graph construction.
Relationship Extraction
Automatic extraction technique of semantic relationships between entities from texts, enabling continuous enrichment of knowledge graphs with new structured information.
Knowledge Graph Embedding
Dense vector representation of entities and relationships from a knowledge graph in a continuous space, facilitating mathematical operations and semantic inferences.
Semantic Reasoning
Logical inference process applied to knowledge graphs to deduce new information from existing facts, thus extending the coverage and depth of retrieved knowledge.