Lecture 10 : Deep Learning for NLP

Question 1

What are 2 things deep learning models use for NLP?

Answer 1

-Vector representation of words–>word embeddings
-Neural network structures

Question 2

What is Word2Vec?

Answer 2

-Popular embedding method
-Very fast to train
-Idea: predict rather than count + use texts from the Web(unsupervised learning)

Question 3

What are the 2 training methods of Word2Vec?

Answer 3

1.CBOW (continuous bag of words): given context words, guess the word

2.Skip-gram: given a word, guess one of its surrounding words