Techniques & Methods
Entity Extraction
Entity extraction (a component of NER) finds spans of text corresponding to named entities and assigns them to predefined types. It is the foundation of knowledge base population, relationship extraction, and information retrieval systems.
Modern entity extraction systems use fine-tuned BERT-based or generative LLM models that handle complex, ambiguous cases—such as distinguishing "Apple" (company) from "apple" (fruit) using surrounding context.
Authority Links
Related Terms
Techniques & Methods
Named Entity Recognition (NER)
Identifying and classifying named entities in text into predefined categories like people and places.
Core Concepts
Entities
Specific, identifiable elements like names, places, and dates extracted from text.
Techniques & Methods
Entity Annotation
Labeling text spans with entity type information to create structured training data.
Techniques & Methods
Information Extraction
Automatically extracting structured information from unstructured text.