She first ran a dataset of 365 examples, and the results weren’t particularly useful or amusing. However, one of her readers – Jo Scott – helped her out by compiling all 1,300 spells from the 4th edition. “Using the new dataset,” Shane explained, “I was able to train a much better-performing neural network. It simply had many more examples of spells to work with; that is, more examples of the words and letter combinations that appear in D&D spells, and thus was able to deduce better rules about how to create them.”
“Even this more-sophisticated neural network is not without some oddities,” Shane continues. “For example, you’ll notice in the results below that it seems to have a particular fondness for bears. And it has invented the name ‘Dave’ which is now, for some reason, its favorite.”
I leave you now with just a few of the wonderful Dungeons & Dragons spells generated by the latest neural network. For the full list of results, including all the Daves, check out Shane’s Lewis & Quark blog.
Tag yourselves, I’m Curse Clam.
Next up, Shane hopes to use her neural network to generate Dungeons & Dragons character names. You can help her with this new project by entering your D&D character’s name, race, and class via this form. As with the spells, the more data she has on hand, the better the results will be – so let’s help her out!
(Via Lewis and Quark; image via Wizards of the Coast, LLC)
Want more stories like this? Become a subscriber and support the site!
—The Mary Sue has a strict comment policy that forbids, but is not limited to, personal insults toward anyone, hate speech, and trolling.—