Uses examples of sequential music and text generation to show how and when recurrent neural networks (specifically, long short-term memory networks) are useful.