Deep Learning Applications: Natural Language Processing, Part 2

This lecture is about how language tokens, such as words, are represented in deep learning. We begin with the distinction between conventional symbolic representation of words and the distributed representation used by neural networks. Several "word-to-vec" methods are described.