Relation Mention Detection Task

A Relation Mention Detection Task is a Mention Detection Task that requires the Detection of the Relation Mentions within a Text Document.

• AKA: Semantic Relation Mention Detection Task, Semantic Relation Detection from Unstructured Data Task, Co-occurrence Relation Mention Detection Task.
• Context:
  • Input:
    • a Text Corpus.
    • Optionally, the Entity Mention Types in the relation. E.g. Relation(Protein,Protein).
  • output:
    • The set of Relation Mentions contained in the corpus
    • Optionally, a Relation Mention Detection Model.
  • Performance:
    • Correctness (e.g. Precision)
    • Resource Consumption (e.g. Computational Complexity).
  • TypeOf: It is a type of Relation Detection Task.
  • It can be solved automatically by a Relation Mention Detection System (that implements a Relation Mention Detection Algorithm) or manually by a Domain Expert.
  • It can support a Relation Mention Recognition Task.
  • The Relation(s) can be specified to be of the type Domain Dependent Relation or Domain Independent Relation.
• Example(s):
  • "A cat is a mammal" ⇒ Relation(cat,mammal)
  • "My cookie has chocolate chips." ⇒ Relation(chocolate chips, cookie).
  • "Alexander went to Australia." ⇒ Relation(Alexander, Australia).
  • "Microsoft is based in Redmond." ⇒ Relation(Microsoft, Redmond).
  • "XyaA is one of E. coli’s proteins. It is found in the periplasm." ⇒ Relation(E. coli, XyaA), Relation(E.coli, periplasm), Relation(XyaA, periplasm).
  • "Albert's niece, Ann, got engaged to John." Relation(Albert,Ann) ^ Relation(Ann, John).
  • SemEval-1 Task 4 (Benchmark Task))
  • Message Understanding Conference (Benchmark Task))
  • ACE Benchmark Task.
  • PPLRE Benchmark Task.
• See: Relation Mention Classification Task; Entity Mention Detection Task.

• SemEval-2007. http://nlp.cs.swarthmore.edu/semeval/program.php

• Concept