When we hear about Convolutional Neural Network (CNNs), we typically think of Computer Vision. CNNs were responsible for major breakthroughs in Image Classification and are the core of most Computer Vision systems today, from Facebook’s automated photo tagging to self-driving cars.
More recently we’ve also started to apply CNNs to problems in Natural Language Processing and gotten some interesting results. In this post I’ll try to summarize what CNNs are, and how they’re used in NLP. The intuitions behind CNNs are somewhat easier to understand for the Computer Vision use case, so I’ll start there, and then slowly move towards NLP.
Opinion phrases was published on September 02, 2014 and last modified onSeptember 02, 2014 by Vlad Sandulescu.
Website and Github:
In Predicting what user reviews are about with LDA and gensim I played with extracting topics from short reviews and given a new review, tried to predict the most probable topic(s) it can be associated with. LDA relies on a bag-of-words model, which is a very popular document representation approach. The model disregards any syntactic dependencies between the words, i.e. any grammar, as well as word order in the documents. For a deeper read about the assumptions made by the LDA model, try to digest Blei’s paper…if you dare!