Shallow Semantic Parcing is a system for identifying the semantic relationship, or semantic roles, filled by constituents of a sentence within a semantic frame. Given an input sentence and a target word and frame, the system labels constituents with either abstract semantic roles or more domain-specific semantic roles.
We can find some problem with the Shallow Semantic Parcing:
- “Argument Identification”: some semantic arguments in the predicate can´t be identifyed.
- “Argument Classification”: it´s difficult to assing the appropiate argument labels to the given constituents known to represent arguments of a predicate.
- “Argument Identification and Classification”: It is a combination of the above two.
References:
- Shallow Semantic Parcing. In Stanford University. Retrieved 21 June 2009, 11:00. From http://nlp.stanford.edu/projects/shallow-parsing.shtml
- Automatic Labeling of Semantic Roles. In Stanford University. Retrieved 21 June 2009, 11:30. From http://www.stanford.edu/~jurafsky/cl01.pdf
- Shallow Semantic Parsing uing Support Vector Machines. In Stanford University. Retrieves 21 June 2009, 11:52. From http://www.stanford.edu/~jurafsky/hlt-2004-verb.pdf
