Paper Link
Authors: Michael Gutmann
Aim:
This Paper is second in the series by Mikolov et al. of Statistical Language Models Based on Neural Networks
It is an extension to the proposed Skip-gram model in terms of quality and training speed. It also incorporates methods for Phrase representation and explores the
interesting property of Skip-gram model.
Notes on Noise Contrastive Estimation and Negative Sampling
comments powered by Disqus