1. (37 marks) Submit your answer either as plain text in one of the options: a file a4q3.txt, PDF file named a4q3.pdf, an image file named a4q3.jpg, or a few image files called a4q3-1.jpg, a4q3-2.jpg, etc.

Consider the following treebank of bracketed parse trees:

(S (VP (VB define)(NN cosmology))(E .))

(S (VP (VB define)(NNP thalassemia))(E .))

(S (VP (VB name)(NP (DT a)(NN stimulant)))(E .))

(S (SBARQ (WHNP (WP what))(SQ (VBP are)(NP (NNPS amphibians)))(E ?)) )

(S (SBARQ (WHNP (WP what))(SQ (VBP are)(NP (NNS antacids)))(E ?)) )

(S (SBARQ (WHNP (WP what))(SQ (VBZ is)(NP (NNP java)))(E ?)) )

(S (SBARQ (WHNP (WP who))(SQ (VP (VBD discovered)(NP (NN electricity))))(E ?)) )

(S (SBARQ (WHNP (WP what))(SQ (VBZ does)(NP (NNP lol))(VP (NN mean)))(E ?)) )

Each line represents one parse tree of a sentence that is a request to an Question-Answering (QA) system. Each sentence is either a direct request or a question.  All terminal symbols are lower-case words or punctuation at the end of the sentence.

• (3 marks) What are terminal symbols in this treebank?
  1. (4 marks) What are non-terminal symbols (i.e., variables) in this treebank? What is the start symbol?
  1. (10 marks) What are the context-free rules induced from this small treebank? With these rules, and parts a) and b) we can see the full CFG induced by this treebank.
  1. (5 marks) Is the grammar obtained in a) in CNF (Chomsky Normal Form) or not? Explain your answer.
  1. (10 marks) What is the PCFG that is learned from this treebank? You must use the rule counting as discussed in the class; i.e., maximum likelihood learning.
  • (5 marks) Write the left-most derivation in which at every step you choose the most likely rule of all rule choices. If there is a tie between more than one rule, choose the rule which comes first alphabetically according to their left-hand side.

• (30 marks) Submit your answer either as plain text in one of the options: a file a4q4.txt, PDF file named a4q4.pdf, and image file named a4q4.jpg, or a few image files called a4q4-1.jpg, a4q4-2.jpg, etc.

Consider the following sentences:

You must submit your source files with images. and

You must submit your source files with C code.

The sentences should have different syntactic structure because the first sentence means that you must submit your source files and your image files, while the second sentence means that you must submit your source files, which are actually C source files.

This difference can be clearly shown with two parse trees for the sentences.

1. (10 marks) Draw the parse trees for two sentences using the standard tags and rules.
2. (10 marks) Annotate the parse trees with the head information and dependencies (as done in the example in class).
3. (10 marks) Write both parse trees using the bracketed notation.

Get expert help for Treebank of bracketed parse trees and many more. 24X7 help, plag free solution. Order online now!