irb intro

Description of Project:

Word2Vec4Kids is a MacOS application that is designed to teach children about word embeddings to further their understanding of Artificial Intelligence. Users of the application will first learn basics about how computers interpret text and then interact with various game modes to facilitate a deeper understanding.

Data Collection:

The only information a user will be asked to provide is their first name and grade level. Word2Vec4Kids will automatically store data that pertains to specific modes and tie that data to the user’s first name and grade. In the first mode, called quiz, we ask the user a multiple-choice question. For this mode we will store what question was asked by the application as well as all of the selections that the student makes before going to the next quiz question and how long it took to get to the right answer. In the second mode, called challenge, the user is presented with a starting word and an ending word. We will measure the length of time it takes to get to the final word, along with the selection of words chosen to get to the final, as there is a *reset* button we will be recording the number of times that button is selected and how many times hints were used. In the third mode, playground, we give the user a blank canvas to add and subtract words to create new ones. For this mode we will store the number of words they create, the chain of words to their final word, and how long each session took.

Post Assessment Questionnaire

1.     How do computers interpret text?

a.     Like us, as text

b.     With numbers

2.     What advantages, if any, does an embedding like word2vec have over the binary number representation?

3.     Machine learning models know…

a.     Everything

b.     Almost everything

c.     what they’ve seen

d.     very specific information

4.     How are embeddings made?

a.     Searching over common location like twitter/Wikipedia

b.     Creating new text, like writing a book

c.     Add more numbers to a binary representation

d.     Taking in all known text

5.     I would explain how computers see and understand text to a friend by saying:

6.     I would describe Word2Vec4Kids to my friends/family as: