Articles → NLP → Word2vec (CBOW Version) In NLP

Word2vec (CBOW Version) In NLP

Purpose

1. Input → Context words
2. Output → Target word (centre word).

Steps

1. Tokenization
2. Choose window size
3. Create CBOW training pairs
4. Convert words to vectors (one-hot encoding)
5. Passing data to a neural network

Tokenization

I love working with AI

[I, love, working, with, AI]

Choose Window Size

Create CBOW Training Pairs

Convert Words To Vectors (One-Hot Encoding)

I       → [1,0,0,0,0]love    → [0,1,0,0,0]working → [0,0,1,0,0]with    → [0,0,0,1,0]AI      → [0,0,0,0,1]

Passing Data To A Neural Network

Target = "working"Context = ["love", "with"]

love  → [0,1,0,0,0]with  → [0,0,0,1,0]

([0,1,0,0,0] + [0,0,0,1,0]) / 2 = [0, 0.5, 0, 0.5, 0]

working → [0,0,1,0,0]