DataTerminology.comNEW
Also known as: Named Entity Disambiguation
Mapping text mentions to canonical entities in a knowledge base, resolving ambiguity using context and external knowledge.
DESCRIPTION
Entity linking connects mentions in text to unique knowledge base entries like Wikipedia or Wikidata. It improves search, question answering, and knowledge graph building by grounding references. Systems combine mention detection, candidate generation, and disambiguation using context, embeddings, and priors. Handling NIL cases where no entity matches is a key challenge.
ALTERNATIVE DEFINITIONS
Wikipedia
Entity linking (EL), also known as named entity disambiguation (NED), is the task of assigning a unique identity to entities mentioned in text.
Defines EL/NED and emphasizes assigning unique identities.
Stanford NLP
Entity linking is the process of mapping textual mentions to entries in a knowledge base, resolving ambiguity using context.
Stresses mapping mentions to knowledge base entries with contextual disambiguation.
BLINK (Facebook AI)
Entity linking aims at aligning entity mentions in text with the corresponding entities in a reference knowledge base.
Research framing focusing on alignment with reference KBs.
RELATED TERMS
Knowledge Base Named Entity Recognition Coreference Resolution Wikidata Knowledge Graph Disambiguation Knowledge Representation NIL Detection
COMPETING TERMS
KNOWLEDGE AREAS
AI Data Science Semantic & Knowledge Modeling
TARGET AUDIENCE
Data Scientist AI Engineer Data Engineer Data Steward Information Analyst
EXTERNAL SOURCES
- Wikipedia - Entity linking
- DBpedia Spotlight - Open-source EL system documentation
- Facebook Research - Entity linking model repository
- Wikidata - Open knowledge base for linking
TOP QUESTIONS
Why is Entity Linking important for knowledge graphs?
It grounds textual mentions to canonical entities, enabling consistent graph construction and queryability.
Category: Relevancy
What are the key steps in an Entity Linking pipeline?
Mention detection, candidate generation, and disambiguation using context and priors.
Category: Requirements
What is NIL detection in Entity Linking?
It identifies mentions with no matching entity in the knowledge base, crucial for handling out-of-KB references.
Category: Topic
What mistakes occur in Entity Linking systems?
Over-reliance on popularity priors and insufficient context windows lead to biased or incorrect links.
Category: Mistakes
What risks arise from poor Entity Linking?
Misinformation propagation, incorrect analytics, and degraded search relevance are possible.
Category: Risks
What competes conceptually with Entity Linking?
Entity resolution links records across datasets; it overlaps but addresses a different problem setting.
Category: Alternatives
How do LLMs impact Entity Linking?
LLMs improve contextual understanding but require grounding to avoid hallucinations in linking.
Category: Opinions
What’s a practical example of Entity Linking?
Mapping ‘Paris’ in text to the city entity in Wikidata rather than the person or brand.
Category: Example
Where to start learning Entity Linking?
Explore DBpedia Spotlight, BLINK, and Stanford NLP chapters on linking and KBs.
Category: Learning