Back to Subject

Unit 3: Natural Language Processing with Probabilistic Models - Practice Quiz

INT344 — Natural Language Processing 50 Questions
0 Correct 0 Wrong 50 Left
0/50

1 What is the primary goal of calculating Minimum Edit Distance?

A. To quantify the dissimilarity between two strings by counting operations
B. To generate word embeddings for a neural network
C. To calculate the probability of a word appearing in a sentence
D. To find the longest common subsequence between two strings

2 Which of the following operations is NOT typically used in the Levenshtein distance algorithm?

A. Substitution
B. Transposition
C. Deletion
D. Insertion

3 In the context of dynamic programming for edit distance, if source[i] equals target[j], what is the cost of substitution?

A. 0
B. Infinity
C. 1
D. 2

4 What is the Minimum Edit Distance between the strings 'cat' and 'cut'?

A. 0
B. 1
C. 2
D. 3

5 How does an autocorrect system typically identify candidate words for a misspelled word?

A. By finding words within a certain edit distance threshold
B. By selecting random words from the dictionary
C. By using the longest word in the dictionary
D. By choosing words that start with the same letter only

6 Which algorithm is commonly used to efficiently calculate the Minimum Edit Distance?

A. Depth-First Search
B. Gradient Descent
C. K-Means Clustering
D. Dynamic Programming

7 In Part of Speech (POS) tagging, what is the 'Hidden' component in a Hidden Markov Model?

A. The punctuation marks
B. The transition probabilities
C. The Part of Speech tags
D. The words in the sentence

8 What does the Markov Assumption state in the context of Markov Chains?

A. The future state is independent of the current state
B. The future state depends only on the current state
C. The future state depends on the hidden emissions
D. The future state depends on all past states

9 What are 'Transition Probabilities' in an HMM?

A. The probability of moving from one POS tag to another
B. The probability of generating a specific word given a tag
C. The probability of a sentence starting with a specific word
D. The probability of a word being misspelled

10 What are 'Emission Probabilities' in an HMM used for POS tagging?

A. P(tag|previous_tag)
B. P(tag|word)
C. P(word|tag)
D. P(word|previous_word)

11 Which algorithm is used to find the most likely sequence of hidden states (POS tags) given a sequence of observations?

A. The Edit Distance Algorithm
B. The Backward Algorithm
C. The Viterbi Algorithm
D. The Forward Algorithm

12 A text corpus is:

A. A large, structured set of texts used for statistical analysis
B. A dictionary of word definitions
C. A set of rules for grammar
D. A software used for autocorrection

13 In an N-gram language model, what is 'N'?

A. The number of hidden states
B. The dimension of the word embedding
C. The number of words in the sequence considered for probability
D. The number of words in the sentence

14 Which assumption simplifies the calculation of N-gram probabilities?

A. The probability depends on the entire sentence history
B. All words have equal probability
C. The probability of a word depends only on the previous N-1 words
D. Words are independent of each other

15 How is the probability of a bigram P(w2 | w1) calculated from a corpus?

A. Count(w1) * Count(w2)
B. Count(w1, w2) / Count(w2)
C. Count(w1, w2) / Count(w1)
D. Count(w1) / Count(w2)

16 What is the main problem with N-gram models when N is very large?

A. The context window becomes too small
B. Data sparsity (many sequences have zero counts)
C. The model becomes too simple
D. The vocabulary size decreases

17 What technique is used to handle N-grams that have zero probability in the training data?

A. Filtering
B. Smoothing (e.g., Laplace Smoothing)
C. Pruning
D. Tagging

18 In Laplace (Add-1) smoothing, what is added to the denominator?

A. 1
B. The vocabulary size (V)
C. The number of sentences
D. The total word count

19 What does an autocomplete system try to maximize?

A. P(previous_words | next_word)
B. The edit distance between words
C. The length of the sentence
D. P(next_word | previous_words)

20 A Trigram model looks at how many previous words to predict the next word?

A. 3
B. 0
C. 1
D. 2

21 One-hot encoding of words results in vectors that are:

A. Dense and low-dimensional
B. Dense and high-dimensional
C. Sparse and low-dimensional
D. Sparse and high-dimensional

22 What is a major limitation of one-hot encoding for words?

A. It does not capture semantic similarity between words
B. It requires a neural network
C. It is difficult to compute
D. It cannot represent rare words

23 Word embeddings typically represent words as:

A. Strings
B. Sparse binary vectors
C. Integers
D. Dense vectors of real numbers

24 Which metric is commonly used to measure the similarity between two word embedding vectors?

A. Jaccard Index
B. Cosine Similarity
C. Perplexity
D. Edit Distance

25 The Word2Vec model 'Skip-gram' architecture tries to predict:

A. The next sentence
B. The POS tag of the word
C. The target word given the context words
D. The context words given the target word

26 The Word2Vec model 'CBOW' (Continuous Bag of Words) architecture tries to predict:

A. The part of speech
B. The document topic
C. The target word given the context words
D. The context words given the target word

27 What famous algebraic property is often cited to demonstrate the semantic capability of word embeddings?

A. Fast + Slow = Speed
B. Paris - France = Germany
C. Apple + Orange = Fruit
D. King - Man + Woman = Queen

28 In the 'Noisy Channel Model' for spelling correction, P(x|w) represents:

A. The probability of x being a valid word
B. The probability that the user meant w but typed x
C. The probability of typing x given the intended word w
D. The probability of the word w appearing in the corpus

29 When building an HMM for POS tagging, the sum of probabilities of all outgoing transitions from a single state must equal:

A. 0
B. 1
C. The number of observations
D. The number of states

30 What is 'Perplexity' in the context of Language Models?

A. The time taken to train the model
B. A measure of how well a probability model predicts a sample
C. The number of parameters in the model
D. The size of the vocabulary

31 Why do we use Log Probabilities instead of raw probabilities in N-gram calculations?

A. To handle negative numbers
B. To increase perplexity
C. To make numbers larger
D. To avoid arithmetic underflow

32 Which of the following describes the 'Start' token (<s>) in N-gram models?

A. It represents an unknown word
B. It gives context for the first word in the sentence
C. It indicates the end of a sentence
D. It is used for punctuation

33 What represents the 'Observations' in a POS HMM?

A. The transition matrix
B. The sequence of tags
C. The initial state probabilities
D. The sequence of words in the text

34 In Minimum Edit Distance, the 'backtrace' step is used to:

A. Calculate the cost
B. Determine the actual sequence of operations (alignment)
C. Initialize the matrix
D. Sum the rows

35 Which token is typically used to replace words not found in the training vocabulary?

A. <NULL>
B. <END>
C. <START>
D. <UNK>

36 A 'Unigram' model assumes that:

A. Words depend on the grammar
B. Words depend on the previous word
C. Words are independent of context
D. Words depend on the previous two words

37 The dimensionality of a Word2Vec embedding vector is typically chosen by:

A. The length of the sentence
B. The system designer (hyperparameter)
C. The number of unique characters
D. The size of the vocabulary

38 In the equation P(tag|word) ∝ P(word|tag) * P(tag), what is P(tag)?

A. Posterior probability
B. Likelihood
C. Emission probability
D. Prior probability

39 If we want to build a spell checker, which probability do we want to maximize according to Bayes' theorem?

A. P(correction | typo)
B. P(correction)
C. P(typo | correction)
D. P(typo)

40 Which type of language model suffers most from the 'curse of dimensionality'?

A. Bag of Words
B. High-order N-gram model (e.g., 5-gram)
C. Unigram model
D. Word2Vec

41 What is the primary input to a neural network training a Word2Vec model?

A. Image pixels
B. Audio signals
C. One-hot encoded vectors of words
D. Parse trees

42 The term 'corpus' in NLP refers to:

A. A specific neural network layer
B. A body of text data
C. A computer algorithm
D. A type of spelling error

43 In edit distance, if we assign a higher cost to substitution than insertion/deletion, it implies:

A. The algorithm will fail
B. Insertion is impossible
C. The distance will always be zero
D. Typing a wrong letter is considered worse than missing a letter

44 What is the result of using a sliding window in N-gram generation?

A. It removes stop words
B. It creates a sequence of overlapping word chunks
C. It calculates the edit distance
D. It converts text to uppercase

45 Which of these words likely has a vector closest to 'frog' in a well-trained embedding space?

A. Toad
B. Steel
C. Philosophy
D. Galaxy

46 In an HMM, what connects hidden states to each other?

A. The Viterbi path
B. Observation vectors
C. Emission probabilities
D. Transition probabilities

47 What is 'Stupid Backoff' in the context of Language Models?

A. A type of neural network
B. A way to delete wrong words
C. A method to stop the algorithm
D. A smoothing method that uses lower-order N-grams if higher-order ones are missing

48 Which application primarily utilizes Probabilistic Language Models?

A. Image Compression
B. Speech Recognition
C. Network Routing
D. Database Management

49 In the context of Word Embeddings, what does 'Polysemy' refer to?

A. Words that rhyme
B. Words in different languages
C. Words with similar spellings
D. Words with multiple meanings

50 If P(A|B) is the probability of tag A following tag B, this is an example of:

A. Observation probability
B. Edit probability
C. Emission probability
D. Transition probability
Previous MCQ
Unit 2
Next MCQ
Unit 4