About this document

Enhancing Word Representations with Skip-gram by ujnzaq is a document available to read on EtoBox.

This paper presents extensions to the Skip-gram model for learning word embeddings from large amounts of text. It introduces subsampling of frequent words to speed up training and improve representations of less frequent words. It also describes negative sampling, a simpler alternative to hierarchical softmax that results in faster training and better representations of frequent words. Additionally, it presents a method for learning embeddings of phrases by treating phrases as individual tokens, which impro

Author

ujnzaq

Language

EN