Google translates your voice into text, Facebook suggests that the man wearing a hat and sunglasses is your friend Mike, and IBM guesses exactly what recipe you'd love to eat for dinner tonight. Deep Learning algorithms are the hands-down most state-of-the-art machine learning algorithms for each of these tasks.
These same Data Science teams have created open-source programming tools to quickly allow any Data Scientist to use these Deep Neural Networks on their own machines (and Cloud) and on their own problems.
This presentation will introduce attendees to the mechanics and concepts of these Neural Network models, demystifying the opaque nature of Deep Learning as a "black box." We will further clarify their description through their application in the world of Natural Language Processing.
We will present various use cases of Deep Nets on text data, ranging from sentiment analysis and opinion mining of Twitter and Yelp review data to the application on internal email and pdf documents. These use cases are industry-agnostic and applicable to widely-varying business teams.
While de-mystifying Deep Learning, attendees will leave with sufficient knowledge and exposure to get started developing Deep Learning models on their own data sets and problems. Attendees will walk away with working code and implementations of Deep Neural Networks on publicly-available data.
All of the tools will come from the open-source community, using Python and Jupyter Notebooks.