Introduction to Deep Learning for Natural Language Processing

Ever wondered how Google Translate works so well? Do you know how the auto-captioning works on Youtube? How to mine the sentiments of tweets on twitter? And there’s a host of successful natural language processing examples out there. What’s the underlying theme? They all use Deep Learning.

This workshop introduces artificial neural networks and deep learning. The building blocks of neural networks are discussed in detail. Attendees are introduced to learning using ANN with backpropagation algorithm.

A preliminary model using multi-layer perceptron is implemented to get a feel for deep learning model structure and the library keras.

The state-of-art recurrent neural networks is introduced. Two approaches are shown - 0ne using Long Short-Term Memory models and another one leveraging Word2Vec. Some ways to overcome overfitting are explored.

We’ll also show how GPU’s affect the computation.

We’ll also implement a RNN to do text-generation.

Ideal Background of the attendee

Some knowledge of machine learning. - Bias and Variance - Train, Validation, and Test sets - Cross-validation - hyperparameter optimization - Measuring model accuracy - Supervised and Unsupervised learning - Overfitting - Basic Python constructs

The repository for the tutorial is here

The repository has instructions on what packages to install and the data we would be using for the workshop.