Technical GlossaryNatural Language Processing

Open Information Extraction

An approach that extracts subject-relation-object structures from text without relying on a predefined relation schema.

Open Information Extraction aims to obtain structured information from text without requiring a fixed ontology or predefined relation set. This flexibility is valuable for exploratory extraction across large document collections. However, the open-schema setup also introduces challenges in quality control and normalization. It is often used as an initial layer in knowledge graph construction.