I’ve used neural networks to name all kinds of things - halloween costumes, craft beers, cats, and even guinea pigs. The weirder the starting set of names, the more a neural network’s creations might blend in (although cats named “Jexley Pickle” and “Big Wiggy Bool” might at least raise eyebrows).
How about the names of racehorses? The Jockey Club recently released a list of over 42,000 registered racehorses “in an effort to assist owners in identifying an appropriate name for their Thoroughbreds.” They don’t say whether it’s okay to use it to identify inappropriate names for Thoroughbreds instead.
So I trained two neural networks, char-rnn and textgenrnn, on this tantalizing dataset. Some of the names they generated might pass for the names of real horses (which after all include horses named Three Little Birds, Spider Magic, Fierce Pajamas, and even a prizewinning horse called Cloud Computing).
She’s a Babe
(I decided to illustrate a few using BigGAN (via ganbreeder.app), another neural net that generates pictures. Unfortunately, “horse” is not a category of image it can do, so I decided to go with “horse cart”. There was an attempt.)
The fact that these names are not taken yet highlights the potential of neural nets for this important task. The next set of names, however, highlights their disadvantages.
Culf Q’s Wart
Aaging C Bon
Warry L Z’ W.
(For these illustrations I experimented with mixing in birds for more colorful horses - hornbill, and flamingo)
If you want your horse to be memorable, though, neural net names may be just the thing.
Killer the Butterfly
Pickle J Gator
Pat’s Quick Hat
Left: “horse cart” plus a bit of “anenome fish”. Right: “horse cart” plus a bit of “great grey owl”. “Dog” crept in there because BigGAN saw a LOT of dogs during training.
AI Weirdness supporters get bonus content: names than would fit here! Or become a free subscriber to get new posts in your inbox.