Using deep learning to generate offensive license plates by Jonathan Nolis.
From the post:
If you’ve been on the internet for long enough you’ve seen quality content generated by deep learning algorithms. This includes algorithms trained on band names, video game titles, and Pokémon. As a data scientist who wants to keep up with modern tends in the field, I figured there would be no better way to learn how to use deep learning myself than to find a fun topic to generate text for. After having the desire to do this, I waited for a year before I found just the right data set to do it,
I happened to stumble on a list of banned license plates in Arizona. This list contains all of the personalized license plates that people requested but were denied by the Arizona Motor Vehicle Division. This dataset contained over 30,000 license plates which makes a great set of text for a deep learning algorithm. I included the data as text in my GitHub repository so other people can use it if they so choose. Unfortunately the data is from 2012, but I have an active Public Records Request to the state of Arizona for an updated list. I highly recommend you look through it, it’s very funny.
…
What a great idea! Not only are you learning deep learning but you are being offensive at the same time. A double-dipper!
A script for banging against your state license registration is left as an exercise for the reader.
A password generator using phonetics to spell offensive phrases for c-suite users would be nice.