Chapter 5. Analyzing Sentiment with a Bidirectional LSTM
This chapter is a bit more practical to get a better sense of the commonly used recurrent neural networks and word embeddings presented in the two previous chapters.
It is also an opportunity to introduce the reader to a new application of deep learning, sentiment analysis, which is another field of Natural Language Processing (NLP). It is a many-to-one scheme, where a variable-length sequence of words has to be assigned to one class. An NLP problem where such a scheme can be used similarly is language detection (english, french, german, italian, and so on).
While the previous chapter demonstrated how to build a recurrent neural network from scratch, this chapter shows how a high-level library built on top of Theano, Keras, can help implement and train the model with prebuilt modules. Thanks to this example, the reader should be able to decide when to use Keras in their projects.
The following points are developed in this chapter...