This week you'll learn about Word2Vec: A Study of Embeddings in NLP.

Image

Last week we delved into the world of representation learning in natural language processing (NLP) with Bag-of-Words. However, due to the looming problems of inefficiency and effectively ignoring the beautiful intricacies of language (namely individual word meanings and grammar), the Bag-of-Words architecture couldn't be hailed as the magnum opus of NLP. 

That brings us to the next attempt at making language discernable for computers, Word2Vec. An ingenious idea constrains the data into finite bounds and presents each word as an understandable tokenized sequence to the computer. But how does this happen?

The big picture: Similar to Bag-of-Words, Word2Vec involves using vectors for representation, but instead of sentences, each word is represented as a vector in an N-dimensional embedding space. 

How it works: The architecture uses associations formed due to the occurrence of words with each other. Hence, the final vector representation for a word will be determined by none other than its neighboring words. 

Our thoughts: Word2Vec is a brilliant jump from Bag-of-Words, essentially trying to figure out where each word fits inside our embedding space. This way, even paradigms like grammar can be ideally retained if occurrences of words are good enough. 

Yes, but: This still wasn't the magnum opus for which the NLP world was waiting. It still had problems of inaccuracies and inefficiencies, but it drove NLP in the right direction. 

Stay smart: Don't forget to try these architectures on your own data! 

Click here to read the full tutorial

Solve Your CV/DL problem this week (or weekend) with our Working Code

You can instantly access all of the code for Word2Vec: A Study of Embeddings in NLP by joining PyImageSearch University. Get working code to

  1. Finish your project this weekend with our code
  2. Solve your thorniest coding problems at work this week to show off your expertise
  3. Publish groundbreaking research without multiple tries at coding the hard parts

Guaranteed Results: If you haven't accomplished your CV/DL goals, let us know within 30 days and get a full refund.

I want the code

Note: You may have missed this, but last Wednesday, we published a new post on Computer Vision and Deep Learning for Customer Service.


The PyImageSearch Team