Term Mention Linking Task

AKA: Term Mention Resolution.
Context:
- Input:
  - a Word Mention String that is restricted to Term Mentions.
  - a Word Sense Inventory that is restricted to Terminological Unit Records (a Technical Term Inventory).
- output:
  - a Sense Tagged Expression that is restricted to Disambiguated Term Mentions.
- It can by supported by a Term Mention Recognition Task.
- It can be solved by a Term Mention Reference Resolution System (that implements a Term Mention Reference Resolution Algorithm.
- It can be a Term Mention Classification Task if only a small set of Term Records are required (e.g. PROTEIN/GENE/ORGANISM).
Example(s):
- a Wikipedia-based Term Mention Resolution Task.
- …
Counter-Example(s):
- Term Mention Recognition Task.
See: Term Mention Coreference Resolution Task, Reference Resolution Task, Word Sense Disambiguation Task.

References

(Krauthammer & Nenadic, 2004) ⇒ Michael Krauthammer, and Goran Nenadic. (2004). “Term Identification in the Biomedical Literature.” In: Journal of Biomedical Informatics, 37(6). doi:10.1016/j.jbi.2004.08.004
- We differentiate three main steps for the successful identification of terms from literature: term recognition, term classification, and term mapping. ... While classification helps to establish some broad notion of the nature of a biomedical concept, it is not sufficient for establishing term identity. This is done by term mapping, which links terms to well-defined concepts of referent data sources, such as controlled vocabularies or databases. The linking definitely establishes the exact term identity (with respect to the referent data source). Mapped terms are annotated with referent identifiers (IDs) that act as keys to supplementary information such as preferred and synonymous terms, or sequence information. The mapping of terms is essential in any data integration efforts where acquired knowledge on specific biomedical concepts is aggregated across different data sources.