Use There Was an Old Lady Who Swallowed a Fly to answer the questions, below. Note that the title of the song is a part of the corpus!
Text Generation Exercise #1
1 Let’s start our text generation with the phrase "She ".
Text Generation 1: Shefirst word second word third word fourth word
2 Determine which word is most likely to follow "She " and record it on the line for second word above.
3 Determine which word is most likely to follow the word you just wrote down and record it on the line for third word above.
4 Use statistical language modeling to determine the fourth word.
5 Everyone in your class should have generated the same text. Why do you think that was the outcome?
Text Generation Exercise #2
Here is a list of the most common unigrams in the corpus: "a": 11 times "she": 11 times "the": 12 times
6 Let’s start by choosing the most common word:
7 Determine which word is most likely to follow that word:
8 There are two words that have an equal probability of appearing in the third spot! What are they? and
9 Flip a coin to determine which one you will use to complete Text Generation 2a and use statistical language modeling to determine the fourth word. Then repeat the process for Text Generation 2b using the other word that tied for the third spot.
Some students will have the 2a response in the 2b blanks, and vice versa — which is fine!
Text Generation 2a: thefirst word second word third word fourth word
Text Generation 2b: thefirst word second word third word fourth word
10 Why was there only one result for Text Generation Exercise 1, while Text Generation Exercise 2 had two possible results?
★ What is another starting word that would result in only one possible four-word phrase?
★ ★ What is another starting word that would result in at least two possible outcomes?
These materials were developed partly through support of the National Science Foundation, (awards 1042210, 1535276, 1648684, 1738598, 2031479, and 1501927).
Bootstrap by the Bootstrap Community is licensed under a Creative Commons 4.0 Unported License. This license does not grant permission to run training or professional development. Offering training or professional development with materials substantially derived from Bootstrap must be approved in writing by a Bootstrap Director. Permissions beyond the scope of this license, such as to run training, may be available by contacting contact@BootstrapWorld.org.