Let’s say you have a dataset of Things. Things could be people, cities, countries, dinosaurs, episodes of The Simpsons, whatever really. You’d like to better understand the dataset as quickly as you can, use it to improve your understanding of Things, and maybe even build software features on top of it. Basically, you want the dataset “at your fingertips”.

With modern specialized computing power, neural networks that generate audio are more commonplace. Training these with backpropagation requires a loss function that can take two audio representations – a model’s current best guess and the true target sound – and compute a similarity score with differentiable functions.

Drums libraries have been one of the most important developments in digital music production. We’re no longer restricted to the stock sounds that come with drum machines, allowing for more room to develop a unique sound. We’ve seen a surge in free drum kits on the web, and there’s an ever-growing list of digital audio workstations with which to use them.