Word2Vec
Linear Digressions
00:00
Is This a Binary Classifier?
Instead of trying to predict which word it is, of all the valid options, what word to ve does is it makes the model a binary classifier. So you don't have to compare each word to the ten thousand other words in the language. You just have to compare it to a smaller number of noise words. And it allows you to re formulate it as a binary classification problem also,. Is this a real word, or is this a noise word? As opposed do they're all real words, and you have to like, pick the best one?, i see. I think they're just collections of characters, although i didn't get into this detail as much, so i could be
Play episode from 10:25
Transcript


