We've created a few months ago a french Youtube channel "Le Machine Learning par l'exemple" where people could discover descriptions of recent and powerful machine learning algorithms in a series of 5 minutes videos called Algoflash.
Word2vec is a Nautral Language Processing algorithm introduced by T. Mikolov et al at NIPS 2013. It transforms words into vectors by representing words in terms of its context. The main idea behind word2vec is to predict the best representation of a given word by observing its nearby words. It uses Stochastic Gradient Descent optimization of a likelihood based on a logistic model. Word2vec has shown impressive results for different NLP tasks and proposes a meaningfull representation of words in a high dimensional space. Some insights are explained in the 5' video here.
We just sent you an email. Please click the link in the email to confirm your subscription!