Quizz 02: General Intro to NLP + Language Models

Name: ________________________

1. Consider the various definitions of the linguistic unit called "word": orthographic (sequence of letters separated by delimiters); semantic (an independent unit of meaning); dictionary (a unit which is listed in the lexicon of the language). For each of the following cases, indicate whether the classification of the string between brackets as a single word according to the [O]rthographic, [S]emantic [D]ictionary sense:

   
   He even [wouldn't] agree to eat bread.
   
   [It's] on the right.
   
   [Trump's] tweet triggered a sworm of protest.
   
   [Come on]!
   
   [Beer Sheva] has more fountains than Tel Aviv.
   
   [Two hundred and fifty five] students attended the event
   

2. Define what is a language model

   
   
   
   
   

3. What are the two probabilistic models involved in spell checking, and how are they combined?

   
   
   
   
   

4. Write the chain-rule expression of the joint probability p(w1,...,wn)

   
   
   
   
   
   
   

5. What is the Markov assumption applied to language modeling

   
   
   
   
   
   
   

6. Define the perplexity of a bigram language model given a dataset [w1...wN]

   
   
   
   
   
   
   

7. How many parameters are needed to model a bigram language model for a vocabulary of size V?