Course Code: 23CS0540 R23

SIDDHARTH INSTITUTE OF ENGINEERING & TECHNOLOGY:: PUTTUR

(AUTONOMOUS)
Siddharth Nagar, Narayanavanam Road – 517583
BIT BANK (OBJECTIVE)

Subject with Code: NATURAL LANGUAGE PROCESSING (23CS0540)

Course & Branch: [Link] – CSM & CAI
Regulation: R23 Year & Sem: III - [Link] & I-Sem
UNIT - 1
Introduction to Natural language

1 Which of the following is an application of NLP? [ ]

a) Machine Translation b) Sentiment Analysis c) Chatbots d) All

2 The study of how computers understand human languages is called: [ ]

a) Machine Translation b) Natural Language Processing c) Compiler Design d) Artificial Vision

3 Which level of language analysis deals with the structure of words? [ ]

a) Morphological level b) Syntactic level c) Pragmatic level d) Semantic level

4 Which level of analysis is concerned with the meaning of sentences? [ ]

a) Pragmatics b) Syntax c) Semantics d) Morphology

5 In Natural Language Processing, Syntax refers to: [ ]

a) Structure of words b) Structure of sentences c) Meaning of words d) language in context

6 Pragmatics in Natural Language Processing studies: [ ]

a) Sentence grammar only b) Meaning of isolated words c) Language use in context d) Speech signals

7 Phonology in Natural Language Processing refers to: [ ]

a) Sentence meaning b) Sound system of language c) Sentence structure d) Word formation

8 The organization of NLP systems typically includes: [ ]

a) Input, Processing, Output b) Rules, Learning, Decision c) Data, Storage, Retrieval d) None

9 Which is a linguistic background component in NLP? [ ]

a) Syntax b) Semantics c)Pragmatics d) All of the above

10 The first chatbot ELIZA was developed in: [ ]

a) 1956 b) 1966 c) 1946 d)1986

11 Which component converts speech into text? [ ]

a) Speech recognition b) Syntax parser c) Semantic analyzer d) Text generator

12 Which is a knowledge-based NLP approach? [ ]

a) Rule-based b) Statistical c) Machine learning d) Neural networks

Course Code: 23CS0540 R23
13 A statistical Natural Language Processing system relies on: [ ]

a) Grammar rules b) Probabilities and data c) Only semantics d) Hardware design

14 In Natural Language Processing, ambiguity means: [ ]

a) One sentence has many meanings b) Sentence has no meaning c) Wrong grammar d) No structure

15 Which is NOT an NLP application? [ ]

a) Sentiment analysis b) Compiler optimization c) Question answeringd) Machine translation

16 Knowledge representation in NLP means: [ ]

a) Storing language data b) Representing meaning of words c) Storing images d) Writing programs

17 Bag-of-Words model is used for: [ ]

a) Image processing b) Text representation c) Speech generation d) Grammar rules

18 Word embeddings represent words as: [ ]

a) Rules b) Symbols only c) Syntax trees d) Vectors

19 Which model represents text as “term frequencies”? [ ]

a) ATN b) Parse tree c) TF-IDF d) All of the above

20 Which representation ignores word order? [ ]

a) Parse tree b) Semantic net c) Bag-of-Words d) CFG

21 Google Translate is an example of: [ ]

a) Compiler b) Operating system c) Search Engine only d)Machine Translation

22 Siri and Alexa are based on: [ ]

a) Only databases b) Operating systems c) Only compilers d) NLP+Speech recognition

23 In Natural Language Processing, Text classification is used in: [ ]

a) Spam detection b) Disk scheduling c) CPU scheduling d) All of the above

24 In Natural Language Processing, Sentiment analysis checks: [ ]

a) Emotions/Opinions in text b) Grammar errors c) Speech volume d) Memory usage

25 Which is a Natural Language Processing task? [ ]

a) Page replacement b) Binary addition c) Part-of-Speech tagging d) Sorting numbers

26 In Natural Language Processing, Syntax deals with: [ ]

a) Sound recognition b) Word meaning c) Context of use d) Sentence structure

Course Code: 23CS0540 R23
27 In NLP, Parsing is the process of: [ ]

a) Storing files b) Translating words only c) Generating speech sounds d) Analyzing sentence
structure

28 The general english syntax is usually: [ ]

a) Subject-Verb-Object b) Object-Verb-Subject c) Verb-Subject-Object d) Verb-Object-Subject

29 In Natural Language Processing, A parse tree shows: [ ]

a) Meaning of a word b) Structure of a sentence c) Emotions in text d) Frequency of terms

30 CFG in NLP stands for: [ ]

a) Central File Generator b) Context-Free Grammar c) Context File Graph d) None of the above

31 Evaluating NLP systems means checking: [ ]

a) Hardware design b) Compiler errors c) Accuracy of understanding d) Disk usage

32 In NLP, BLEU score is used for: [ ]

a) Compiler testing b) CPU scheduling c) File management d) Machine translation evaluation

33 The Precision in Natural Language Processing measures: [ ]

a) Correct results among retrieved ones b) All retrieved results c) Speed of algorithm d) Grammar
errors

34 The Recall in Natural Language Processing measures: [ ]

a) Algorithm speed b) Grammar checking c) Correct results retrieved from all relevant d) Wrong results
only

35 The F1 score in Natural Language Processing measures: [ ]

a) Average accuracy b) Harmonic mean of Precision & Recall c) Syntax error rate d) All of the
above

36 The term Tokenization in Natural Language Processing means: [ ]

a) Splitting text into words/tokens b) Removing words in a line c) Translating sentences d) Combining
words

37 In Natural Language Processing, The Stop words are: [ ]

a) Rare words b) Named entities c) Common words ignored in NLP d) Technical terms

38 In Natural Language Processing, POS tagging means: [ ]

a) Removing words b) Removing stop words c) Splitting sentences d) Assigning part of speech to
words

39 Which is an open-source NLP library? [ ]

a) NLTK b) Photoshop c) Oracle d) MS Excel

Course Code: 23CS0540 R23
40 Which is NOT a challenge in NLP? [ ]

a) Context understanding b) Word sense disambiguation c) Ambiguity d) File allocation methods

Course Code: 23CS0540 R23
UNIT - 2
Grammars and Parsing

1 In Natural Language Processing, a grammar defines: [ ]

a) Word meaning b) Sentence structure c) Speech sound d) Context

2 Which grammar type is widely used for describing programming languages and natural languages in NLP?
[ ]

a) Regular Grammar b) Context-Free Grammar (CFG) c) Dependency Grammar d) Feature Gramma

3 In Natural Language Processing, Parsing is the process of: ]

a) Removing stop words b) Translating text c) Tokenizing words d) Analyzing sentence structure

4 Which part of a grammar specifies the basic vocabulary (words or tokens)? [ ]

a) Parse trees b) Productions c) Terminals d) Non-terminals

5 Which is an example of a non-terminal in grammar? [ ]

a) Sentence (S) b) "dog" c) "run" d) "the"

6 In Natural Language Processing, Top-down parsing starts from: [ ]

a) Start symbol(root) b) Words c) Tokens d) Morphemes

7 In Natural Language Processing, Bottom-up parsing starts from: [ ]

a) Word meaning b) Grammar rules c) Root node d) Words

8 The main purpose of parsing is to: [ ]

a) Checking sentence follows grammatical rules b) Remove stop words c) Morphemes d) None

9 Which type of parsing analyzes a sentence from the start symbol to the words? [ ]

a) Bottom-up parsing b) Top-down parsing c) Statistical parsing d) Dependency parsing

10 Shift-Reduce parsing is an example of: [ ]

a) Bottom-up parsing b) Top-down parsing c) Semantic analysis d) Feature analysis

11 In Natural Language Processing, ATN stands for: [ ]

a) Advanced Token Network b) Augmented Transition Network c) Automated Tree Node d) all the
above

12 In Natural Language Processing, an ATNs extend: [ ]

a) Morphological rules b) CFGs c) Parse trees d) Finite state machines

13 In Natural Language Processing, Features in grammars capture:: [ ]

a) Word meaning b) Stop words c) Properties like number or tense d) Word frequency
Course Code: 23CS0540 R23
14 What is the example of a feature system: [ ]

a) TF-IDF b) Subject-verb agreement c) Word embeddings d) N-grams

15 In Natural Language Processing, Augmented grammar adds: [ ]

a) Embeddings b) Tokens c) Features and constraints d) Stop words

16 The smallest meaningful unit of language is called: [ ]

a) Morpheme b) Token c) Word d) Sentence

17 A lexicon in NLP is best described as: [ ]

a) Word meaning b) A dictionary that stores words and properties c) Speech sound d) Context

18 Which is an example of derivational morphology? [ ]

a) "walk" → "walked" b) "cat" → "cats" c) "run" → "running" d) "happy" → "happiness"

19 Which is an example of inflectional morphology? [ ]

a) "beauty" → "beautiful" b) "teach" → "teacher" c) "run" → "runs" d) "use" → "useful"

20 In Natural Language Processing, Ambiguity is: [ ]

a) Simple solution b) Major challenge c) Rare case d) Ignored case

21 Entropy in Natural Language Processing measures: [ ]

a) uncertainty in predicting words b) length of a sentence c) no of parse trees d) The frequency of token

22 Bayes’ Rule in NLP is used for: [ ]

a) Stop word removal b) Tokenization c) Probabilistic parsing d) Word embeddings

23 Bayes’ Rule formula can be written as: ]

a) none of the given b) P(A|B) = P(A) × P(B) c) P(A|B) = P(A) + P(B) d) P(A|B) = [P(B|A) × P(A)]
/ P(B)

24 The Shannon game in NLP is related to: [ ]

a) Predicting the next word b) Splitting text into tokens c) Removing stop words d) Context

25 In Natural Language Processing, Cross-entropy is used to:: [ ]

a) identify grammar rules b) Build parse trees c) Count tokens in text d) Predicted vs actual
distribution

26 which of the following Feature-based parsing ensures: [ ]

a) Agreement constraints b) No ambiguity c) Tokenization d) Stemming

27 what is Structural ambiguity means: [ ]

Course Code: 23CS0540 R23
a) Word has many meanings b) Multiple sentence structures c) Wrong grammar d) Extra punctuation

28 what is Lexical ambiguity means: [ ]

a) No stop words b) Sentence with multiple trees c) Word with multiple meanings d) Extra
morphemes

29 what is PCFG stands for: [ ]

a) ProbabilisticContext-Free Grammar b) Parsed CFG c) Programming CFG d) Phrase Context

Grammar

30 what PCFG helps by in NLP: [ ]

a) Assigning probabilities to rules b) Removing tokens c) Adding embeddings d) Adding embeddings

31 In Natural Language Processing, Chart parsing uses: [ ]

a) Lexicons only b) Recursion only c) Tokens only d) Dynamic programming

32 In NLP, CKY parsing is used for : [ ]

a) Tokenization b) Dynamic programming for CFGs c) Speech sound d) Context

33 Shift-reduce conflict happens due to: [ ]

a) Lexicons b) Tokens c) Ambiguity d) Stop words

34 Reduce-reduce conflict happens when: [ ]

a) No tokens left b) Stop words appear c) Two reductions possible d) Wrong grammar rules

35 Parsing efficiency improves with: [ ]

a) Dynamic programming b) Stop words c) Extra tokens d) Random guesses

36 The Term Recursive descent parser is: [ ]

a) Feature parser b) Statistical parser c) Bottom-up parser d) Simple top-down parser

37 In Natural Language Processing, Statistical parsers resolve: [ ]

a) Ambiguity using probability b) Tokens into morphemes c) Speech sound d) Context words

38 How Finite-state automata are used in: [ ]

a) Neural networks b) Morphological analysis c) Semantic parsing d) Word embeddings

39 Parsing in NLP is important for: [ ]

a) File allocation b) Stop word removal c) Syntax analysis d) Data compression

40 Which parser is most suitable for large-scale NLP applications with ambiguous sentences?
[ ]
a) Greedy parser b) Simple rule-based parser c) Naïve parser d) Probabilistic parser
Course Code: 23CS0540 R23
UNIT - 3
Grammars for Natural Language

1 In NLP, grammars for natural languages are more complex than those for programming languages because
they deal with: [ ]

a) Fixed rules b) Human ambiguity c) Limited syntax d) Binary data

2 Which grammar type is commonly used to describe natural language sentence structures? [ ]

a) Context-Free Grammar b) Regular Expression c) Automata Grammar d) Symbol Grammar

3 The main challenge of natural language grammars is handling: [ ]

a) Strict rules b) Ambiguity c) Compilation d) Hardware

4 Unlike programming languages, natural languages are: [ ]

a) Deterministic b) Ambiguous c) Unambiguous d) Binary

5 Which approach uses dependency between words to describe grammar? [ ]

a) Phrase structure b) Statistical grammar c) Dependency grammar d) Regular grammar

6 In English, the movement phenomenon is often observed in: [ ]

a) Exclamations b) Commands c) Statements d) Questions

7 The movement of a word or phrase from its original position is known as: [ ]

a) Movement b) Displacement c) Substitution d) Reduction

8 Which type of movement involves auxiliary verbs moving to sentence-initial position? [ ]

a) Verb raising b) Shifting c) Inversion d) Swapping

9 Example: “What did you see?” shows movement of which word category? [ ]

a) Pronoun b) Wh-word c) Noun d) Adverb

10 The phenomenon where parts of sentences appear in different positions while preserving meaning is called:
[ ]

a) Reordering b) Movement c) Displacement d) Shuffling

11 In grammar, a “gap” refers to: [ ]

a) Missing element b) Extra phrase c) Wrong word d) Duplicate token

12 The technique used to connect gaps to their original positions is called: [ ]

a) Linking b) Threading c) Mapping d) Parsing

13 Gap threading is useful in analyzing: [ ]

a) Simple words b) Long-distance dependencies c) Tokens only d) Syntax errors

Course Code: 23CS0540 R23
14 Example: “The book that I read __ was amazing.” The blank is a: [ ]

a) Rule b) Token c) Phrase d) Gap

15 Gap threading helps parsers handle: [ ]

a) Stop words b) Displaced constituents c) Fixed phrases d) Word embeddings

16 Human readers prefer parses that are: [ ]

a) Ambiguous b) More complex c) Simpler d) Random

17 Garden-path sentences demonstrate that humans initially choose: [ ]

a) No parse b) Random parse c) Right parse d) Wrong parse

18 Example: “The old man the boats” confuses readers due to: [ ]

a) Lexical choice b) Garden-path effect c) Semantic clarity d) Word order

19 Human parsing preference is often influenced by: [ ]

a) Meaning b) Syntax c) Frequency d) Grammar

20 Human preferences in parsing are also shaped by: [ ]

a) Probability b) Experience c) Memory d) Context

21 A shift-reduce parser works by shifting input onto a stack and: [ ]

a) Expanding b) Reducing c) Reversing d) Matching

22 The “shift” operation means: [ ]

a) Push token onto stack b) Pop token from stack c) Replace non-terminal d) End parsing

23 The “reduce” operation means: [ ]

a) Replace symbols on stack b) Add tokens to stack c) Delete sentence d) Stop parsing

24 Shift-reduce parsing is an example of: [ ]

a) Top-down b) Bottom-up c) Hybrid d) Probabilistic

25 A common problem in shift-reduce parsing is: [ ]

a) Conflicts b) Ambiguity c) Memory d) Time

26 A deterministic parser tries to build parse trees with: [ ]

a) Multiple choices b) Ambiguous trees c) Random rules d) Single choice

27 A parser that does not backtrack and makes one decision at each step is called: [ ]

a) Ambiguous b) Non-deterministic c) Deterministic d) Recursive

Course Code: 23CS0540 R23
28 Deterministic parsers work best with: [ ]

a) Hidden grammars b) Ambiguous grammars c) Random grammars d) Unambiguous grammars

29 The main advantage of deterministic parsers is: [ ]

a) Efficiency b) Ambiguity c) Randomness d) Slowness

30 The disadvantage of deterministic parsing is: [ ]

a) Limited grammar handling b) Infinite recursion c) Random results d) No parsing

31 Grammars for natural language often combine: [ ]

a) Syntax and hardware b) Syntax and semantics c) Tokens and machines d) Rules and compilers

32 Phrase structure grammars represent: [ ]

a) Probabilities b) Word embeddings c) Hierarchical sentence structure d) Stop words

33 Human parsing differs from machine parsing because humans rely on: [ ]

a) Probability only b) Grammar only c) Rules only d) Experience

34 Dependency grammars represent: [ ]

a) Relations b/w words b) Relations b/w sentences c) Relations b/w documents d) Relations b/w sounds

35 Ambiguity in grammar can be reduced using: [ ]

a) Complier b) Hardware design c) Probabilistic methods d) Tokenization

36 Which parser type handles movement phenomena better? [ ]

a) ATN b) Recursive descent c) Deterministic d) Greedy

37 Long-distance dependencies are especially challenging for: [ ]

a) PCFG b) CFG c) ATN d) HMM

38 Probabilistic parsing improves accuracy by using: [ ]

a) Frequencies b) Stop words c) Syntax trees d) Morphemes

39 A shift-reduce conflict occurs when the parser: [ ]

a) Cannot tokenize b) Must choose b/w shift and reduce c) Loses input d) Ignores grammar

40 Deterministic parsing is faster but fails when grammar is: [ ]

a) Fixed b) Simple c) Linear d) Ambiguous

Course Code: 23CS0540 R23
UNIT - 4
Semantic Interpretation and Language Modelling

1 In semantic interpretation, the goal is to convert syntactic structures of sentences into a form that represents
their: [ ]

a) Meaning b) Syntax c) Tokens d) Rules

2 A logical form is used in NLP to: [ ]

a) Represent sentence in logic b) Encode only phonetic sounds c) Remove ambiguity d) Convert text to
tokens

3 Word sense ambiguity occurs when: [ ]

a) word has multiple meanings b) grammar rule is missing c) parser cannot tokenize d) sentence is
incomplete

4 Thematic roles such as Agent, Theme, and Location are used in: [ ]

a) Syntax rules b) Semantic role labeling c) Morphological parsing d) Tokenization

5 The study of how sentences express actions, states, or events in logical form is called: [ ]

a) Word segmentation b) Syntax parsing c) Event semantics d) Grammar mapping

6 Which semantic element represents the intention behind a sentence, such as questioning or commanding?
[ ]

a) Parsing rules b) Word embedding c) Tokenization d) Speech acts

7 The method of encoding ambiguity directly into logical forms is known as: [ ]

a) Tokenization b) Underspecification c) Reduction d) Simplification

8 In semantic analysis, mapping natural language sentences to logical formulas is part of: [ ]

a) Formal semantics b) Morphology c) Dependency parsing d) Feature grammar

9 The model-theoretic approach in semantics focuses on: [ ]

a) Lexical entries b) Grammar rules c) Truth conditions d) Token lists

10 A system that interprets word senses in context to resolve ambiguity is performing: [ ]

a) WSD b) POS tagging c) Parsing d) Clustering

Course Code: 23CS0540 R23
11 The problem of determining which sense of a word is activated in a sentence is called: [ ]

a) WSD b) Parsing c) Tokenization d) Lemmatization

12 When the same sentence can be interpreted in multiple structures, it represents: [ ]

a) Syntactic ambiguity b) Word segmentation c) Semantic noise d) Token repetition

13 A word like “bank” meaning either “river side” or “financial institution” is an example of: [ ]

a) Lexical ambiguity b) Phonological variation c) Morphological agreement d) Statistical grammar

14 The process of mapping synonyms or related words into the same meaning space is part of: [ ]

a) Lexical semantics b) POS tagging c) Parsing rules d) statistical models

15 Polysemy in semantics refers to: [ ]

a) Multiple related meanings b) Unrelated meanings c) Morphological agreement d) Statistical grammar

16 Homonymy refers to words that share spelling but differ in: [ ]

a) Grammer b) Syntax c) Meaning d) Structure

17 Which ambiguity occurs when context cannot decide between two or more word senses? [ ]

a) Lexical b) Statistical c) Phonetic d) Morphological

18 A semantic parser is mainly designed to resolve: [ ]

a) Parsing depth b) Ambiguity c) Stop words d) Word order

19 Which ambiguity arises from words like “light” meaning “not heavy” or “illumination”? [ ]

a) Lexical b) Structural c) Syntactic d) Phonemic

20 Which semantic concept assigns meaning by linking words to knowledge bases like WordNet?? [ ]

a) Lexicon b) Automaton c) Parser d) Tokenizer

21 In NLP, a language model is a system that: [ ]

a) Predicts word sequences b) Stores syntax rules c) Classifies phonemes d) Builds parse trees

22 The probability of a sentence in language modelling is often approximated using: [ ]

a) N-grams b) Tokens c) Stemming d) Parsing

23 A bigram model predicts the next word based on: [ ]

a) Sentence meaning b) Random choice c) Previous word d) Stop word

24 In an n-gram model, larger values of “n” capture: [ ]

a) More context b) No context c) Stemming d) Single words

Course Code: 23CS0540 R23
25 A limitation of n-gram models is their: [ ]

a) Data sparsity b) Grammar depth c) Parsing order d) Lexical redundancy

26 Perplexity is a measure of: [ ]

a) Language model quality b) Word segmentation c) Sentence parsing d) Syntax tree depth

27 Cross-entropy in language modelling measures: [ ]

a) Difference b/w predicted and true distribution b) Sentence length c) Syntax errors d) Morphological
variation

28 Which smoothing technique adjusts zero probabilities in n-gram models? [ ]

a) Laplace b) Clustering c) parsing d) Tokenization

29 The process of adapting a model to a new domain or dataset is called: [ ]

a) Adaptation b) Segmentation c) Annotation d) Stemming

30 Which type of language model is based on neural networks? [ ]

a) Neural LM b) Statistical LM c) Lexical LM d) Hybrid LM

31 A multilingual language model can handle: [ ]

a) Multiple languages b) One language c) Random tokens d) Stop words only

32 Cross-lingual language modelling is useful for: [ ]

a) Translating across languages b) Parsing simple words c) POS tagging only d) Morphological
segmentation

33 A challenge in multilingual modelling is: [ ]

a) Parsing speed b) Token loss c) Grammar overlap d) Data imbalance

34 Code-switching in multilingual text refers to: [ ]

a) Mixing languages b) Mixing phonemes c) Mixing morphemes d) Mixing grammar rules

35 Which evaluation metric is used for language model performance? [ ]

a) Perplexity b) Token count c) Syntax tree d) Grammar depth

36 Neural language models are preferred because they: [ ]

a) Capture context b) Ignore context c) Remove tokens d) Delete syntax

37 Transformer-based models like BERT are trained using: [ ]

a) Masked language modelling b) Morphological grammar c) Parsing rules d) Token splitting

38 Which type of model allows handling unseen words better? [ ]

Course Code: 23CS0540 R23
a) Subword model b) Grammar model c) Parse tree model d) POS model

39 Which evaluation compares system output with human translations or references? [ ]

a) Tokenization b) Recall c) Parsing d) BLEU

40 A major advantage of pre-trained multilingual models is: [ ]

a) Transfer learning b) Stop word removal c) POS tagging d) Tokenization

Course Code: 23CS0540 R23
UNIT - 5
Machine Translation, Multilingual IR & Summarization

1 Machine translation refers to the process of: [ ]

a) convert text b/w languages b) binary into text c) Translate compilers into code d) grammar into parsing

2 The earliest approaches to machine translation were based on: [ ]

a) Statistical models b) Neural networks c) Rule-based systems d) Hybrid techniques

3 A major limitation of rule-based machine translation is its: [ ]

a) Probabilistic strength b) Scalability issues c) Neural accuracy d) Domain adaptation

4 The idea of using probabilities of word sequences for translation led to: [ ]

a) Direct MT b) Neural MT c) Hybrid MT d) Statistical MT

5 Neural machine translation is superior because it captures: [ ]

a) Long-range dependencies b) Handcrafted rules c) Only word frequency d) POS model

6 The Anusaraka system developed in India is designed to: [ ]

a) Provide access b/w Indian languages b) Translate binary c) Support English d) Build neural embedding

7 Neural machine translation uses: [ ]

a) Syntax rewriting b) Hand-coded rules c) Direct word tables d) Deep neural networks

8 Phrase-based translation models improve accuracy by: [ ]

a) Removing syntax b) Ignoring context c) Translating phrases and words d) Using only rules

9 Word alignment in MT refers to: [ ]

a) Mapping b/w source & target b) Parsing grammar trees c) Matching phonetics d) Tokenizing text

10 A key challenge in MT is: [ ]

a) Compiler errors b) Handling ambiguity c) Disk space d) Token length

11 BLEU score is used to: [ ]

a) Train neural networks b) Evaluate translation quality c) Parse grammar d) Remove stop words

12 Higher BLEU score means: [ ]

a) Better translation quality b) Poorer translation quality c) More ambiguity d) Random mapping

13 Human evaluation of MT often considers: [ ]

a) Phonetics b) Syntax only c) Fluency and adequacy d) Phonetics

14 NIST metric is an extension of: [ ]

Course Code: 23CS0540 R23
a) Edit distance b) b) TF-IDF c) BLEU score d) ROUGE score

15 METEOR metric focuses on: [ ]

a) Grammar parsing b) Binary codes c) BLEU score d) Synonyms and recall

16 Multilingual IR is about: [ ]

a) Retrieving documents across languages b) Parsing grammar c) BLEU score d) ROUGE score

17 A common challenge in multilingual IR is: [ ]

a) Disk space b) Compiler errors c) Vocabulary mismatch d) Phonetics

18 A key step in IR is document: [ ]

a) Preprocessing b) Compiling c) Encoding d) Translating

19 Tools like Apache Lucene and ElasticSearch are widely used for: [ ]

a) Information retrieval b) Parsing c) Morphology d) Syntax analysis

20 A multilingual corpus is essential for: [ ]

a) Training and evaluating IR systems b) CPU scheduling c) Operating system design d) Memory allocation

21 Dictionary-based approach uses: [ ]

a) Bilingual dictionary for translation b) Grammar rules c) Neural networks only d) Syntax analysis

22 Corpus-based approach uses: [ ]

a) Phonetic matching b) Grammar simplification c) Binary translation d) Parallel corpora for training

23 Statistical IR models rely on: [ ]

a) Probability distributions b) Token splitting c) Grammar rewriting d) Stopword

removal

24 Vector space models represent documents as: [ ]

a) Syntax trees b) Term-weight vectors c) Binary codes d) Phonetic graphs

25 Language modelling in IR is used for: [ ]

a) Phonetic analysis b) Grammar parsing c) Query likelihood estimation d) Tokenization

26 Summarization in NLP is: [ ]

a) Producing a shorter version of text b) Parsing grammar c) Binary codes d) Phonetic graphs

27 Extractive summarization works by: [ ]

a) Generating new sentences b) Select key sentences/words c) Remove all stopwords d) Reordering tokens

28 Abstractive summarization works by: [ ]

Course Code: 23CS0540 R23
a) Generate new sentences with same meaning b) Select only keywords c) Remove stopwords d) Grammar
parsing

29 Example: Newspaper headline summarization is: [ ]

a) Grammar-based b) Extractive c) Abstractive d) Token-based

30 Example: Picking most frequent sentences is: [ ]

a) Grammar-based b) Extractive c) Random d) Abstractive

31 Frequency-based method for summarization selects: [ ]

a) Binary codes b) Grammar rules c) Sentences with common terms d) Stopwords only

32 Graph-based method for summarization uses: [ ]

a) Binary matching b) Grammar parsing c) Tokenization d) Sentence connectivity

33 Machine learning summarization models use: [ ]

a) Training data and features b) Grammar parsing c) Binary trees d) Stopword removal

34 Deep learning summarization models use: [ ]

a) Binary translation b) Neural networks c) Token splitting d) Abstractive

35 ROUGE metric is commonly used to: [ ]

a) Evaluate summarization quality b) Parse grammar c) Translate text d) Tokenize

36 A challenge in summarization is: [ ]

a) Preserving meaning while reducing text b) Grammar parsing c) Binary codes d)

Tokenize

37 Which summarization type is closer to human-like summaries? [ ]

a) Abstractive b) Extractive c) Hybrid d) Graphical

38 Multi-document summarization means: [ ]

a) Summarize multiple texts into one b) Summarize one text multiple times c) Grammar simplification d) Token
merging

39 ROUGE metric is commonly used to: [ ]

a) Evaluate summarization quality b) Parse grammar c) Translate text d) Tokenize

40 An application of abstractive summarization is: [ ]

a) Grammar parsing b) Tokenization c) Headline generation d) Stopword removal