wayground logo

Free Printable Worksheets

NEW

Font size

S
M
L
XL
Worksheets

Natural Language Processing Quiz

Total questions: 30

Worksheet time: 15mins

Name
Class
Date
1.

Which of the following best describes the main challenge in Natural Language Understanding (NLU)?

a)

Converting speech to text

b)

POS tagging

c)

Tokenizing sentences

d)

Mapping words to structured representations capturing meaning

2.

Which of the following is not a core phase of Natural Language Processing?

a)

Morphological Analysis

b)

Phoneme segmentation

c)

Syntax Analysis

d)

Pragmatic Analysis

3.

The Noisy Channel Model in NLP is primarily used in:

a)

POS tagging

b)

Word sense disambiguation

c)

Spelling correction and speech recognition

d)

Dependency parsing

4.

The main assumption in the Noisy Channel Model is:

a)

Noise can be removed completely by probabilistic inference

b)

Observed data is generated by corruption of a true intended message

c)

Every channel introduces equal noise

d)

Noise is language independent

5.

Which part-of-speech category contributes most to the syntactic structure of English sentences?

a)

Nouns and verbs

b)

Adjectives and adverbs

c)

Prepositions and conjunctions

d)

Pronouns and determiners

6.

A grammar that generates exactly the sentences of a language is known as:

a)

Descriptive grammar

b)

Formal grammar

c)

Prescriptive grammar

d)

Cognitive grammar

7.

Which of the following is a disadvantage of simple N-gram models?

a)

They are non-parametric

b)

They cannot be smoothed

c)

They are not probabilistic

d)

They suffer from data sparsity and lack of generalization

8.

The purpose of smoothing in language models is to:

a)

Increase accuracy by removing stop words

b)

Reassign probabilities to unseen n-grams

c)

Reduce model size

d)

Improve parsing efficiency

9.

Perplexity of a language model is:

a)

The log of word likelihood

b)

The inverse probability of the test set normalized by word count

c)

The measure of ambiguity

d)

The ratio of observed to expected frequencies

10.

Which of the following is not true for neural language models?

a)

They map words to continuous vector spaces

b)

They can capture long-range dependencies

c)

They require fewer parameters than N-grams

d)

They can generalize across similar contexts

11.

A key limitation of feed-forward neural language models is:

a)

They cannot model word sequences of variable length

b)

They cannot process bigrams

c)

They cannot use embeddings

d)

They are unsupervised

12.

Which neural architecture most effectively replaces the N-gram assumption?

a)

RNN-based models

b)

CNN-based models

c)

Naïve Bayes models

d)

CRF-based models

13.

POS tagging involves assigning:

a)

Lemmas to words

b)

Grammatical categories to words

c)

Parse trees to sentences

d)

Dependency arcs between words

14.

Rule-based POS tagging relies primarily on:

a)

Statistical probabilities

b)

Lexical rules and contextual constraints

c)

Machine learning

d)

Embedding spaces

15.

The Hidden Markov Model assumes that:

a)

Words depend on their POS tags

b)

Tags depend only on the previous tag

c)

Words depend on previous words

d)

Tags depend on all previous tags

16.

In an HMM-based POS tagger, emission probabilities represent:

a)

Transition between tags

b)

Probability of word given a tag

c)

Probability of tag given word

d)

Probability of a sequence

17.

A common issue in neural POS tagging is:

a)

Sparse matrix representation

b)

Overfitting on rare words

c)

Lack of parallelism

d)

Fixed sequence length

18.

The major advantage of neural POS tagging over rule-based systems is:

a)

Reduced interpretability

b)

Automatic feature learning

c)

Requirement of large rule sets

d)

Dependency on lexicons

19.

Bottom-up parsing differs from top-down parsing by:

a)

Constructing parse tree from input words upward

b)

Using probabilistic inference

c)

Needing grammar probabilities

d)

Being deterministic

20.

The CKY algorithm is applicable only to grammars in:

a)

Regular form

b)

Chomsky Normal Form

c)

Dependency form

d)

Context-sensitive form

21.

A “treebank” is used for:

a)

Storing lexical data

b)

Annotated corpora with syntactic parse trees

c)

Neural embeddings

d)

Sentence segmentation

22.

The sum of probabilities of all production rules expanding a non-terminal in PCFG must be:

a)

Zero

b)

Less than one

c)

Exactly one

d)

Greater than one

23.

Probabilistic CKY parsing extends CKY by:

a)

Adding backtracking

b)

Multiplying probabilities along parse paths

c)

Removing ambiguity

d)

Adding syntax trees manually

24.

Which parsing approach is most suitable for ambiguous sentences?

a)

Deterministic parsing

b)

Probabilistic parsing

c)

Transition-based parsing

d)

Left-corner parsing

25.

Which of the following is not a limitation of CFGs?

a)

Inability to capture long-distance dependencies

b)

Inability to model hierarchical structure

c)

Inability to represent recursion

d)

Inability to handle ambiguity

26.

Vector semantics is based on which linguistic hypothesis?

a)

Rule-based grammar

b)

Distributional hypothesis

c)

Dependency grammar

d)

Chomsky's hierarchy

27.

In vector space models, similarity between words is most commonly measured using:

a)

Euclidean distance

b)

Cosine similarity

c)

Jaccard coefficient

d)

Pearson correlation

28.

Singular Value Decomposition (SVD) is used in LSA to:

a)

Reduce dimensionality of term-document matrices

b)

Increase vocabulary size

c)

Normalize probabilities

d)

Create new features manually

29.

The concept of “word sense” refers to:

a)

POS category

b)

Context-dependent meaning variations of a word

c)

Morphological form

d)

Speech phoneme

30.

Dense vector embeddings are preferred over one-hot vectors because they:

a)

Are binary

b)

Capture similarity and reduce dimensionality

c)

Are sparse

d)

Represent words independently