Natural Language Processing Parsing and Syntactic Analysis 3

Question 1

The Penn Treebank (PTB) uses which annotation scheme?

Accepted Answer

Phrase-structure (constituency) trees with bracket notation

Answer

The Penn Treebank annotates sentences with bracketed phrase-structure (constituency) trees, defining the gold standard for English parsing.

Question 2

What does the Universal Dependencies (UD) project provide?

Accepted Answer

A cross-linguistically consistent dependency annotation scheme and treebanks for 100+ languages

Answer

Universal Dependencies offers a unified set of dependency relation labels and treebanks across more than 100 languages, enabling cross-lingual NLP research.

Question 3

What is 'garden path' ambiguity in parsing?

Accepted Answer

A sentence that initially seems to follow one parse but requires reanalysis partway through

Answer

Garden path sentences (e.g., 'The horse raced past the barn fell') lead the parser down an incorrect initial interpretation that must be abandoned upon reading more of the sentence.

Question 4

In the context of neural dependency parsing, what does a 'transition-based' parser do?

Accepted Answer

It predicts a sequence of actions (shift, left-arc, right-arc) to build the dependency tree incrementally

Answer

Transition-based parsers build the dependency tree step by step by predicting parsing actions (shift/reduce/arc) using a classifier, making them very efficient.

Question 5

What metric is most commonly used to evaluate the accuracy of a dependency parser?

Accepted Answer

Labeled Attachment Score (LAS) and Unlabeled Attachment Score (UAS)

Answer

LAS measures the percentage of tokens where both the head and the dependency label are correct; UAS measures only the head attachment accuracy.

Natural Language Processing Practice Test

Natural Language Processing Practice Test

Natural Language Processing Parsing and Syntactic Analysis 3