stanford NLP第二节 distributed vs distributional

distributional similarity: is a theory about semantics f word meaning the you can describe the meaning of words by as a use theory of meaning, understanding the context in which they appear.

distributional contrast with denotational

distributed contrast with one-hot word vector

 

PS, 必须要刷的文章：

Learning representations by back-propagating errors.  Rumelhart et al

A neural probabilistic language model. Bengio et al

NLP almost from Scratch. Coolobert and Weston

word2vec hahaha

链式求导

以 e 为底的指数函数 f(x)=exp(x) 求导后还是它本身