Entity Mention Segmentation Task

An Entity Mention Segmentation Task is a Mention Detection Task that is restricted to entity mentions.

• AKA: Entity Mention Detection.
• Context:
  • Input: One or more Documents.
  • output: A set of Entity Mention Locations per Document.
  • Performance Metric: Detection Performance Metrics (Accuracy, Precision, Recall, F-Measure).
  • It can Detect a Unnested Entity Mention or a Nested Entity Mention.
  • It can support an Entity Mention Recognition Task.
  • It can restrict itself to some Noun Type, such as a Proper Noun, or Pronoun.
  • …
• Example(s):
  • a Named Entity Mention Detection Task.
  • a Nominal Entity Mention Detection Task, such as:
    • EMDT("A cat is a mammal.”) ⇒ "A [cat] is a [mammal]."
    • EMDT("A unicycle has one wheel.”) ⇒ "A [unicycle] has one [wheel]."
  • a Pronoun Mention Detection Task.
  • …
• Counter-Example(s):
  • a Semantic Relation Detection Task.
  • an Entity Mention Classification Task.
  • a Visual Entity Segmentation Task.
• See: Entity Mention, Detection Task.

---