I’ve been on a machine learning kick lately. Given a large enough dataset to train with, it’s really interesting to see what a neural network can come up with.
This week, it’s names for craft beer.
If you’re a fan of IPA beer, you’ve got names like Dang River, Yamquak, Yall in Wool, Wicked Geee, Yampy, and Oarahe Momnila Day Revenge Bass Cornationn Yerve Of Aterid Ale. Like strong pale ales? Trippel Lock, Third Maus, Third Danger, Spore of Gold and Drammnt. Stouts more your thing? Look for Sir Coffee, Shock Slate, Take Bean, Black Sink Stout, Shrump, Avidberry, or Cherry Trout Stout.
Naturally, I tried to create my own model using a Python library called Keras and a dataset of 7,500 craft beer names.
…I should leave this stuff to the professionals.
Update: Kaggle has a new tutorial teaching you how to do this exact same thing. Neat!