CSCI 587 - Sample Test 2
NLP in Prolog
- Show how to extend the simple prolog grammar in nlp to handle
- pronouns: e.g., he
- proper names: e.g., Sam
- conjunctions: e.g., Sam Mary and Sue
Parse Trees in Prolog
- Add parse tree(feature) arguments for problem 1.
Bottom-Up Chart Parsing
- Provide a grammar and then show the chart obtained from parsing
"John kissed Mary in the garden."
Features
- AGR,
- VFORM, ROOT, SUBCAT
- COMPFORM, PASS,
- WH, GAP
- SEM, VAR
Parsing with Features
- Explain how to incorporate features into chart parsing.
Parsing Auxiliaries
- Extend the prolog parser to handle the parsing of
the auxillaries: can, should, etc.
- Extend the prolog to handle Subject Aux Inversions?
Parsing wh-Questions
No questions on parsing Wh-questions or gaps for this test.
Semantics
- What is a discourse variable? Also, give an example.
Semantics: Thematic Roles
- The verb feature SUBCAT_np_np generally would require what
thematic roles in addition to the AGENT?
- P259 #3, #5
Compositional Semantics
- Represent the predicate likes(john, sue) as a lambda expression.
- p291 #1, #2
- What phrase would the following SEMs represent?
- (NAME j1 john)
- (PRO i1 IT1)
- < THE h1 HAMMER >
- (LAMBDA x (HIT1 e8 x (NAME j1 john)))
URL = http://sourgum.cs.sc.edu/~matthews/Courses/587/Lectures/SampleTest2.html