NLP MCQ ON MODULE TEST Natural Language Processing

NLP MCQ ON MODULE TEST

Get Natural Language Processing for Free MCQs all at one place for your last-moment preparation. Cdac NLP MCQ Question Modern NLP algorithms are based on machine learning, especially statistical machine learning. The paradigm of machine learning is different from that of most prior attempts at language processing.

NLP MCQ: Improve your understanding of the fundamentals, and algorithms, ... You can build a machine learning RSS reader in less than 30 minutes using -.

 

1.         Which of the following includes major tasks of NLP?

*

1/1

a) Automatic Summarization

b) Discourse Analysis

c) Machine Translation

d) All of the mentioned

 

 

2.         Choose from the following areas where NLP can be useful.

*

1/1

a) Automatic Text Summarization

b) Automatic Question-Answering Systems

c) Information Retrieval

d) All of the mentioned

 

 

3.         What are the main challenge/s of NLP?

*

1/1

a) Handling Ambiguity of Sentences

 

b) Handling Tokenization

c) Handling POS-Tagging

d) All of the mentioned

 

4.         The Bag-of-Words approach_________

*

1/1

1) keeps word order, keeps word multiplicity

2) keeps word order, disregards word multiplicity

3) disregards word order, disregards word multiplicity

4) disregards word order, keeps word multiplicity

 

 

5.         Which step is the process of breaking down documents into smaller units of analysis?
*

1/1

1. Ngrams

2. Tokenization

 

3. Stopwords

4. Corpus

 

6.         Which is a model of measuring the incidence of known words?
*

1/1

1. A low weight in TF-IDF

2. A high weight in TF-IDF

3. A corpus

4. A bag of words

 

 

7.         From the sentence “Fintech Online Course”, how many bigrams can be created?
*

1/1

1. 2

 

2. 3

3. 4

4. 1

 

8.         Which are included in named entity recognition?
*

1/1

1. Currency

2. Time and dates

3. Nouns

4. All of above

 

 

9.         Which is a collection of documents?
*

1/1

1. Corpus

 

2. Ngrams

3. Tokenization

4. Stopwords

 

10.       What is the role of Stemming and Lemmatization in NLP?

*

1/1

A) To summarize a long text into a shorter version

B) To translate text from one language to another

C) To reduce words to their base or root form to help with processing and analysis

 

D) To identify the named entities in a text

 

11.       What are word embeddings used for in NLP (Natural Language Processing)?

*

1/1

A) To convert words into a numerical representation

 

B) To identify the most common words in a text corpus

C) To convert text into images

D) To translate between different human languages

 

12.       What is the main advantage of using word embeddings in NLP models? 

*

1/1

A) They allow for a more interpretable representation of words

B) They reduce the dimensionality of the input data

C) They can capture the context in which words are used

D) All of the above

 

 

13.       Which of the following is true about the vector representations generated by GloVe and word2vec? 

*

1/1

A. GloVe generates sparse representations, while word2vec generates dense representations

B. GloVe generates dense representations, while word2vec generates sparse representations

C. Both GloVe and word2vec generate sparse representations

D. Both GloVe and word2vec generate dense representations

 

 

14.       What is word2vec? 

*

1/1

A. A tool for sentiment analysis

B. A framework for computer vision tasks

C. A word representation model

 

D. A speech recognition software

 

15.       What is the primary use case of Seq2Seq models in NLP? 

*

0/1

A. Sentiment analysis

B. Text classification

C. Machine translation

D. Speech recognition

 

Correct answer

C. Machine translation

 

16.       What is the role of the encoder component in a Seq2Seq model? 

*

1/1

A. To generate the target sequence

B. To process the input sequence and produce a compact representation

 

C. To perform sentiment analysis on the input sequence

D. To classify the input sequence into different categories

 

17.       How does a Seq2Seq model handle the problem of variable-length input and output sequences? 

*

1/1

A. By padding the shorter sequences to match the length of the longest sequence

B. By truncating the longer sequences to match the length of the shortest sequence

C. By training separate models for each different length of input and output sequences

D. By using a recurrent neural network (RNN) with an attention mechanism

 

 

18.       What is the problem of vanishing gradients in RNNs and how is it addressed? 

*

1/1

A. The problem of vanishing gradients occurs when the gradients in the backpropagation process become very small, making it difficult for the network to learn. It is addressed by using a different activation function, such as the rectified linear unit (ReLU).

B. The problem of vanishing gradients occurs when the gradients in the backpropagation process become very small, making it difficult for the network to learn. It is addressed by using a mechanism called Long Short-Term Memory (LSTM).

 

C. The problem of vanishing gradients occurs when the gradients in the backpropagation process become very large, making it difficult for the network to learn. It is addressed by using a different activation function, such as the rectified linear unit (ReLU).

D. The problem of vanishing gradients occurs when the gradients in the backpropagation process become very large, making it difficult for the network to learn. It is addressed by using a mechanism called Long Short-Term Memory (LSTM).

 

19.       How is the output generated in an RNN for NLP tasks? 

*

1/1

A. By using a softmax activation function and treating the output as a probability distribution over the possible classes or words

 

B. By using a sigmoid activation function and treating the output as a binary classification

C. By using a linear activation function and treating the output as a regression problem

D. By using a tanh activation function and treating the output as a continuous value

 

20.       What is the purpose of the output gate in LSTMs? 

*

1/1

A. To determine which information from the cell state to output as the current hidden state

 

B. To determine which information from the previous hidden state to output as the current hidden state

C. To determine which information from the input to output as the current hidden state

D. To determine which information from the cell state and the previous hidden state to output as the current hidden state

 

21.       How does the reset gate in GRUs control the flow of information from the previous hidden 

*

1/1

A. By resetting the hidden state to the initial value

B. By scaling the previous hidden state by a scalar value between 0 and 1

 

C. By adding the previous hidden state to the current hidden state

D. By concatenating the previous hidden state with the current hidden state

 

22.       What is the main difference between LSTM and GRU units in numerical computations? 

*

1/1

A. LSTM has three gates, while GRU has two gates

 

B. GRU is faster to train than LSTM

C. LSTM has a better performance than GRU

D. GRU is easier to implement than LSTM

 

23.       What is Long Short-Term Memory (LSTM)? 

*

1/1

A. A type of activation function

B. A type of pooling layer

C. A type of recurrent neural network

 

D. A type of convolutional neural network

 

24.       What is BERT? 

*

1/1

A. A type of word embedding technique

B. A type of recurrent neural network

C. A type of convolutional neural network

D. A transformer-based pre-trained language model

 

 

25.       What is the main advantage of BERT compared to traditional word embedding techniques? 

*

1/1

A. BERT can capture both the context and the meaning of words

 

B. BERT is faster to train than traditional word embeddings

C. BERT has a better performance on NLP tasks than traditional word embeddings

D. BERT requires less data to train than traditional word embeddings

 

26.       How is BERT used for NLP tasks? 

*

1/1

A. By using the pre-trained BERT model as a fixed feature extractor for the task-specific model

B. By fine-tuning the pre-trained BERT model on a task-specific dataset

 

C. By training a new model from scratch on a task-specific dataset

D. By using the pre-trained BERT model as a language model for generating text

 

27.       What is the main advantage of transformers compared to other types of neural networks in NLP? 

*

1/1

A. They can process sequential data in parallel, reducing the training time

 

B. They can handle long-term dependencies better than RNNs

C. They can handle high-dimensional data better than convolutional neural networks

D. They can handle short-term dependencies better than RNNs

 

28.       What is the purpose of the self-attention mechanism in transformers? 

*

1/1

A. To assign weights to the different parts of the input data based on their relevance to the current task

 

B. To process the input data in parallel, without any sequential dependencies

C. To convert the input data into a fixed-length vector representation

D. To learn a weighted combination of the input data based on their similarity to each other

 

29.       What is the difference between the encoder and decoder in a transformer-based NLP model? 

*

1/1

A. The encoder processes the input data, while the decoder generates the output data

B. The encoder generates the embeddings, while the decoder processes the input data

C. The encoder processes the input data and generates the embeddings, while the decoder generates the output data

 

D. The encoder generates the output data, while the decoder processes the input data

 

30.       What are some common applications of discourse analysis in NLP? 

*

0/1

A. Sentiment analysis and text classification

B. Part-of-Speech (POS) tagging and Named Entity Recognition (NER)

 

C. Machine translation and text summarization

D. Coherence analysis and text generation

Correct answer

D. Coherence analysis and text generation

 

31.       What are some common categories of named entities recognized by NER systems in NLP? 

*

0/1

A. Person names, locations, organizations, and monetary values

B. Verbs, adjectives, and adverbs

C. Numbers and dates

D. Common nouns and proper nouns

 

Correct answer

A. Person names, locations, organizations, and monetary values

 

32.       What is the role of the acoustic model in an ASR system? 

*

1/1

A) To transcribe audio into text

B) To understand the context of speech

C) To model the relationship between audio features and phonemes

 

D) To translate speech into another language

 

33.       What are the primary challenges in ASR systems? 

*

0/1

A) Background noise and accent variations

B) Text generation and summarization

C) Image classification and object recognition

 

D) Speech synthesis and pronunciation

Correct answer

A) Background noise and accent variations

 

34.       How does the attention mechanism work in NLP? 

*

1/1

A) By learning the context of a word in a sentence

 

B) By using statistical methods to predict words

C) By using image recognition algorithms

D) By using neural networks to process speech signals

 

35.       What is the main difference between traditional NLP models and models with attention mechanism? 

*

1/1

A) The use of recurrent neural networks

B) The use of convolutional neural networks

C) The ability to focus on specific parts of the input

 

D) The use of decision trees for classification

 

36.       What is an example of text normalization in NLP? 

*

1/1

A) Converting all characters to lowercase

 

B) Converting all characters to uppercase

C) Removing stop words

D) Keeping the original case of characters

 

37.       What is the main goal of using SVM in NLP? 

*

1/1

A) To classify text into different categories

 

B) To generate text

C) To remove stop words from the text

D) To convert words to their base form

 

38.       What is the most commonly used library for SVM in NLP? 

*

1/1

A) OpenCV

B) Pygame

C) NLTK

D) scikit-learn

 

 

39.       What is the main advantage of using SVM in NLP? 

*

1/1

A) High performance for numerical computing

B) Wide range of libraries for NLP tasks

C) Large community for support and development

D) Ability to handle non-linear and high dimensional data

 

 

40.       Summarization creates new phrases paraphrasing the original source.

*

1/1

a. Extraction-based

b. Abstraction-based

 

C. Auto-correct

d. Procedure based

 

41.       "I Saw The Boy With A Pony Tail ", What Type Of Ambiguity Does Sentence Have

*

1/1

1. Semantic Ambiguity

 

2. Pragmetic Ambiguity

3. Structured Ambiguity

4. Simplex

 

42.       Progmatic Analysis Is the ________ Stage in NLP?

*

1/1

1. First

2. Second

3. Fifth

4. Forth

 

 

43.       Syntactical analysis is done at ______________ level

*

1/1

1. Sentence

 

2. Word

3. Lexicon

4. Symbol

 

44.       In NLP, The process of removing words like “and”, “is”, “a”, “an”, “the” from a sentence is called as

*

1/1

1. Stemming

2. Lemmatization

3. Stop word

 

4. Derivation

 

45.       Famous Stemming algorithm

*

0/1

1. Porter’s Stemmer algorithm

2. Lovins Stemmer

3. Dawson Stemmer

 

4. Snowball Stemmer

Correct answer

1. Porter’s Stemmer algorithm

 

46.       Consider the statement " The students went to class" . Assign POS tags for the statement.

*

1/1

1. DT NN VB P NN

 

2. DT NN NN P NN

3. NN NN VBG P NN

4. DT NN VB P DT

 

47.       What is anaphora?

*

1/1

1. repetition of a phrase at the end of sentences

2. repetition of a phrase at the beginning of sentences

 

3. repetition of a phrase in the middle of sentences

4. repetition of a phrase in a direct row

 

48.       Which of the following is a common method used to determine the similarity between two documents in NLP?

*

1/1

A. Euclidean distance

B. Manhattan distance

C. Jaccard similarity

D. Cosine similarity

 

 

49.       How is cosine similarity calculated? 

*

1/1

A. By dividing the dot product of two vectors by the product of their magnitudes.

 

B. By adding the dot product of two vectors to their product.

C. By multiplying the dot product of two vectors by the product of their magnitudes.

D. By subtracting the dot product of two vectors from their product.

 

50.       What is the range of cosine similarity values? 

*

1/1

A. 0 to 1

B. -1 to 1

 

C. 0 to 2

D. -2 to 2

 

51.       What is the primary data structure used in spaCy for processing text? 

*

1/1

a) List

b) Tuple

c) Dictionary

d) Doc

 

 

52.       How does spaCy handle named entity recognition (NER)? 

*

1/1

a) Using rule-based systems

b) Using machine learning algorithms

c) Using both rule-based systems and machine learning algorithms

 

d) NER is not a feature of spaCy

 

53.       Can spaCy be used for language model training? 

*

1/1

a) Yes

b) No

 

 

54.       What is the default tokenization method used by spaCy? 

*

1/1

a) Word tokenization

 

b) Sentence tokenization

c) Both word and sentence tokenization

d) None of the above

 

55.       Which of the options is not a NLP application?

*

1/1

1. Sentiment analysis

2. Translation from English to French

3. Translation from C++ to Java

 

4. Named entity recognition

 

56.       What is the purpose of negative sampling in the Skip-Gram model?

*

1/1

a) To reduce the computational cost of training the model.

 

b) To reduce the dimensionality of the input data.

c) To increase the accuracy of the model.

d) To increase the interpretability of the model.

 

57.       How does negative sampling work in the Skip-Gram model? 

*

1/1

a) It replaces the softmax function with a sigmoid activation function.

b) It randomly selects a subset of negative examples from the context words.

 

c) It uses a threshold to determine which words are considered positive or negative examples.

d) It increases the size of the hidden layer to improve the model's performance.

 

58. Suppose you have a 10000 word vocabulary, and are learning 500-dimensional word

embeddings.The GloVe model minimizes this objective:

Which of these statements are correct?

*

1/1

a) θi and ej should be initialized to 0 at the beginning of training.

b) θi and ej should be initialized randomly at the beginning of training.

 

 

59.       Which of these equations do you think should hold for a good word embedding? 

*

1/1

A) eboy - egirl = ebrother - esister

B) egirl - eboy= ebrother - esister

 

 

60.       What is CBOW?

*

1/1

a) Continuous bag of words

 

b) Canonical bag of words

c) Constant bag of words

d) All of the above