We'll discuss why the state of the art in NLP has leaped forward in the last 1.5 years. State of the art is well beyond Glove and Word2Vec. Transfer learning is the common approach for modern NLP advancements.
We'll explore how transfer learning works in NLP. Discuss a quick, high level overview of the differences between Google, OpenAI and FastAI approaches to NLP transfer learning. Then we'll show how to apply transfer learning for sentiment analysis classification.
We'll discuss ideas for mapping these concepts onto your problem domains.