Outline:

1. The General structure and assumption of two model: CBOW and Skip-Gram.
2. The simplest version of updating parameters—‘One-word context’ version.
   a. 符号表示及优化目标
   b. 隐藏层到输出层的参数更新
   c. 输入层到隐藏层的参数更新
   d. 感性理解
3. Update parameters version2——‘Multi-word context’ version.
4. Two ways to improve updating parameters
   a. Huffman code & Hierarchical Softmax
   b. Negative sampling

手写笔记（待补充）

Main Reference:

1. 《word2vec Parameter Learning Explained》
2. 《word2vec中的数学》
3. 《word2vec Explained: Deriving Mikolov et al.’s
   Negative-Sampling Word-Embedding Method》

Other Reference:

• The three original paper of Tomas Mikolov:

1. 《Efficient Estimation of Word Representations in Vector Space》
2. 《Distributed-representations-of-words-and-phrases-and-their-compositionality-Paper》
3. 《Distributed Representations of Sentences and Documents》

• Some translation link:
  a. https://www.cnblogs.com/peghoty/p/3857839.html
  b. https://www.cnblogs.com/conan-ai/p/11354926.html

c. https://blog.csdn.net/u010555997/article/details/76598666
d. https://www.jianshu.com/p/4517181ca9c3
e. https://blog.csdn.net/lanyu_01/article/details/80097350