Virtual Labs

What is the main purpose of an N-gram language model?

a: To generate random sentences Explanation

Explanation

b: To estimate the probability of word sequences Explanation

Explanation

c: To translate text between languages Explanation

Explanation

d: To count the number of words in a text Explanation

Explanation

Which of the following is a bigram?

a: the Explanation

Explanation

b: the cat Explanation

Explanation

c: cat sat on Explanation

Explanation

d: sat Explanation

Explanation

Why do we need smoothing in N-gram language models?

a: To make the model run faster Explanation

Explanation

b: To handle zero probabilities for unseen N-grams Explanation

Explanation

c: To increase the vocabulary size Explanation

Explanation

d: To remove stopwords Explanation

Explanation

What does Add-One (Laplace) Smoothing do to the counts in an N-gram model?

a: Subtracts one from each count Explanation

Explanation

b: Adds one to each count Explanation

Explanation

c: Multiplies each count by two Explanation

Explanation

d: Removes all zero counts Explanation

Explanation

If a bigram never appears in the training corpus, what is its probability in a maximum likelihood estimate (without smoothing)?

a: One Explanation

Explanation

b: Zero Explanation

Explanation

c: Depends on the vocabulary size Explanation

Explanation

d: Negative Explanation

Explanation

What is the effect of smoothing on the probability distribution of N-grams?

a: It makes the distribution more uniform Explanation

Explanation

b: It increases the probability of seen N-grams only Explanation

Explanation

c: It decreases the probability of all N-grams to zero Explanation

Explanation

d: It has no effect Explanation

Explanation

Which of the following is a limitation of Add-One Smoothing?

a: It can overestimate the probability of unseen N-grams Explanation

Explanation

b: It requires a very large corpus Explanation

Explanation

c: It cannot be used for bigrams Explanation

Explanation

d: It removes all rare words Explanation

Explanation

How does the vocabulary size (V) affect the denominator in Add-One Smoothing for bigram probabilities?

a: It is added to the count of the previous word Explanation

Explanation

b: It is subtracted from the count of the previous word Explanation

Explanation

c: It is multiplied by the count of the previous word Explanation

Explanation

d: It has no effect Explanation

Explanation

Which of the following is a real-world application of N-gram smoothing?

a: Speech recognition Explanation

Explanation

b: Sorting numbers Explanation

Explanation

c: Image classification Explanation

Explanation

d: Network routing Explanation

Explanation

N-Grams Smoothing