Syntactic Features in NLP Tasks

For our class on 4/13, we will be discussing a range of papers addressing semantically-sensitive NLP tasks using syntactic parse features as a source of information. Each student will read 2-3 different papers. Please come prepared to answer the following questions:

1. What is the task that the paper describes?
2. What parser was used to extract syntactic information?
3. What features did they define over the syntactic structures?
4. Did the features actually help?
5. Was the syntactic information being used as a proxy for semantic information? If so, what was the semantic information they were trying to get at?
6. (If this info is available or inferrable:) In what was did the syntactic structures obscure the semantic information they wanted, and in what ways did it expose it?