Word embedding
Word embedding is a very popular way of representing text data in problems that are solved by deep learning algorithms. Word embedding provides a dense representation of a word filled with floating numbers. The vector dimension varies according to the vocabulary size. It is common to use a word embedding of dimension size 50, 100, 256, 300, and sometimes 1,000. The dimension size is a hyper-parameter that we need to play with during the training phase.
If we are trying to represent a vocabulary of size 20,000 in one-hot representation then we will end up with 20,000 x 20,000 numbers, most of which will be zero. The same vocabulary can be represented in word embedding as 20,000 x dimension size, where the dimension size could ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access