Natural Language Processing - Unit-3
Agenerating all parses
Bchoosing the most likely parse among many possible parses
Cremoving the grammar
Dtranslating the sentence
Correct Answer
choosing the most likely parse among many possible parses
Aselecting the shortest parse
Bselecting a random parse
Crejecting every parse
Dselecting the parse with the highest probability
Correct Answer
selecting the parse with the highest probability
Aremoves all probabilities
Buses no grammar rules
Cannotates non-terminals with their head words
Dworks only for speech
Correct Answer
annotates non-terminals with their head words
Aa second model reorders the n-best parses produced by a first parser
Bthe sentence is reversed
Cthe grammar is deleted
Dthe parser is trained twice on the same sentence
Correct Answer
a second model reorders the n-best parses produced by a first parser
Ause no training data
Bcan only parse speech
Clearn from features to separate correct parses from incorrect ones
Dignore the sentence
Correct Answer
learn from features to separate correct parses from incorrect ones
Atoo few word forms
Bdata sparsity because of the large number of word forms
Cno sentences
Dno syntax at all
Correct Answer
data sparsity because of the large number of word forms
APenn Treebank
BPrague Dependency Treebank
CBrown Corpus
DReuters Corpus
Correct Answer
Prague Dependency Treebank
Afixed constituent order
Bdependency relations
Cmorphology
Dtokenization
Correct Answer
fixed constituent order
Asplitting words into letters
Bmapping natural language text to a formal meaning representation
Ccounting syllables
Dcorrecting spelling
Correct Answer
mapping natural language text to a formal meaning representation
Aimage compression
Bdatabase backup
Cword sense disambiguation, entity resolution and predicate-argument analysis
Dfont selection
Correct Answer
word sense disambiguation, entity resolution and predicate-argument analysis
Adetermining which sense of a word is used in a given context
Bsplitting a word into morphemes
Ctranslating a word
Dcounting word frequencies
Correct Answer
determining which sense of a word is used in a given context
Aa speech corpus
Ba parsing algorithm
Ca lexical database in which words are grouped into synsets
Da neural network
Correct Answer
a lexical database in which words are grouped into synsets
Aa set of antonyms
Ba set of sentences
Ca set of grammar rules
Da set of synonyms that share a meaning
Correct Answer
a set of synonyms that share a meaning
Ameasuring the overlap between dictionary definitions and the context
Bcounting its letters
Cparsing the sentence
Dusing a stemmer
Correct Answer
measuring the overlap between dictionary definitions and the context
Asense-annotated training data
Bno data at all
Conly unlabelled speech
Donly a stop-word list
Correct Answer
sense-annotated training data
Ahas no meaning
Bhas several related senses
Chas only one meaning
Dis always a noun
Correct Answer
has several related senses
Ahave identical meanings
Bhave the same form but unrelated meanings
Chave opposite meanings
Dhave no form
Correct Answer
have the same form but unrelated meanings
Aa more general term, such as 'animal' for 'dog'
Ba more specific term
Can opposite term
Da spelling variant
Correct Answer
a more general term, such as 'animal' for 'dog'
Aa more general term
Ban opposite term
Ca misspelled word
Da more specific term, such as 'dog' for 'animal'
Correct Answer
a more specific term, such as 'dog' for 'animal'
Aonly unlabelled data
Bonly neural embeddings
Chand-written rules
Dno knowledge
Correct Answer
hand-written rules
Ausing sense-labelled corpora only
Btranslating the word
Cclustering the contexts in which a word occurs
Ddeleting ambiguous words
Correct Answer
clustering the contexts in which a word occurs
Aa speech recognizer
Ba parser
Ca stemmer
Da corpus in which words are tagged with WordNet senses
Correct Answer
a corpus in which words are tagged with WordNet senses
Aword sense disambiguation
Btokenization
Cstemming
Dcoreference resolution
Correct Answer
coreference resolution
Aa small amount of labelled data together with a large amount of unlabelled data
Bno data at all
Conly labelled data
Donly rules
Correct Answer
a small amount of labelled data together with a large amount of unlabelled data
Athe colour of the text
Bthe number of pages
Cthe font size
Dword order and morphology
Correct Answer
word order and morphology
Fill in the Blanks
26
WSD stands for Word __________ Disambiguation.
Correct Answer
Sense
27
__________ is a lexical database of English organised into synsets.
Correct Answer
WordNet
28
A set of synonyms that share a meaning is called a __________.
Correct Answer
synset
29
The __________ algorithm disambiguates a word using the overlap between definitions and context.
Correct Answer
Lesk
30
A word that has several related senses shows __________.
Correct Answer
polysemy
31
A hypernym is a more __________ term than its hyponym.
Correct Answer
general
32
A hyponym is a more __________ term than its hypernym.
Correct Answer
specific
33
A PCFG resolves ambiguity by choosing the parse with the highest __________.
Correct Answer
probability
34
A lexicalized PCFG annotates non-terminals with __________ words.
Correct Answer
head
35
Parse __________ reorders the n-best parses produced by a first parser.
Correct Answer
reranking
36
The Prague Dependency Treebank is a treebank for the __________ language.
Correct Answer
Czech
37
Semantic __________ maps natural language text to a formal meaning representation.
Correct Answer
parsing
38
__________ resolution links expressions that refer to the same entity.
Correct Answer
Coreference
39
Languages with many word forms suffer from data __________.
Correct Answer
sparsity
40
WSD that uses no sense-labelled data is called __________ WSD.
Correct Answer
unsupervised