December 4, 1965 – Gemini 7 astronauts Jim Lovell and Frank Borman arrive at the launch pad shortly before blasting off to space.
A neural network learns to create better D&D spells
Neural networks are a type of machine learning program that learns from examples they’re given, rather than relying on a human programmer to invent rules.
In an earlier experiment, I trained a neural network to write new names for Dungeons and Dragons spells based on a list of 365 examples. That’s a really small dataset for a neural network to work with, and I ended up struggling to find training parameters that would strike a balance between word-for-word mimicry of the original list of spells, versus a series of completely made-up words. By filtering extensively through the nonsense, I was able to come up with a short list of interesting new spells. (My favorites were Barking Sphere and Gland Growth).
However, blog reader Jo Scott was kind enough to collect the entire 4th edition list of spells - more than 1,300 spells in all. She explained that she’s playing a character who’s an artificer trying to create an autonomous spellcasting golem - essentially, a magical AI - and she’d like to have more weird spells for the golem to invent. (Her Dungeon Master okayed this and thus only has herself to blame when she has to deal with some of the spells listed below.)
Using the new dataset 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.
For comparison, here’s what the neural network trained on the original spell dataset was producing after it had looked through the spell list 30 times. This is raw, unfiltered output from the neural network.
Original dataset
Wome on frr
Eser Wold
Sereisk
Lelent Warder
Cleater Secfen
Spiritul Plage
Arawen
Speak with Alanc
Plonting Cloud
Aurars
Ensntalice
Stige Dling
Comenthon of Prost
Monsen
Scink
Warrifg
Resser RestractiGn
Cloud
Sreeat
Glasp
Blenss
Bline Ons
Dood to Stone
Aside from a couple of spells that just might work, most of the list is magicky-sounding nonsense, sometimes barely pronounceable.
By contrast, this is what the neural network was producing after it had been trained on the dataset that included all the 4th edition spells:
Full dataset
Curse Word
Crackling claus
Tidal treket
Swirk with
Wall of Storm
Acter Lor distertion
Glib ton
Grasping Mane
Tweel Strike
Revitalizing Strike
Truneming fortune
Fall of the Wild
Tunesrite
Trickstrak empester
Phantasmal assault
Tidalt Atight
Hadabol
Leging Blade
Bund Wind
Dance of Sack
and Prime
Poxsare
Dumination
Mass Cure Fortion
They’re not ALL winners, but the difference is dramatic. This is why, although I can often have fun with small datasets, the really large ones (100,000+ metal bands, or 19,000 IPA beers) tend to produce the most consistently convincing results.
Even this more-sophisticated neural network is not without some oddities. 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 with a selection of Dungeons and Dragons spells generated by the latest neural network.
Mister of Light
Storm of the gifling
Song of goom
Forceful Boor
Chorus of the dave
Maine storm
Frames of Death
Song of the doom goom
Death’s Death’s Proud Bear
Wall of Distraction
Date wards
Plant of Peace
Shield of Farts
Song of the darn
Ward of Snade the Pood Beast
Ice shop
Primal Rear
Summon Storm Bear
Divine Boom
Soul of the bill
Charm of the dave
Spirit of the Spirit
Fire shop
Song of blord
Song of distraction
Forceful Force
Spirit Boating
Song of the ball
Hail to the Dave
Crusading Disk
Summon ass
Call to the Daring
Treeking of Star
Grasping Light
Clinging blade
Primal Prayer Bear
War Cape
Find Strike
Song of the Unworthy
Gate Sail
Icon of Thorns
Song of the door
Star warper
Stone of Death
Chilled arrow
Storm of the dave
Fark Mate
Charm of the cods
Death of the Sun
Greater flick
Curse Clam
Claming Blow
Cursing wink
Conjure Mare
Remorse?
Conjure Bark
Darkworm Colt
Daving fire
Healing of Bat
Mordenkainen’s lucubrabibiboricic angion
There are more of these because I had way too much fun generating spells, but to include them all here would make this post ludicrously long. As usual, you can get the extended list by entering your email at this link (even if you’re already on my mailing list). Just for fun, the extended list isn’t QUITE as filtered to remove all the sweary words.
Also, I thought it would be fun to generate D&D character names for a future project. If you go to this form (no email required), you can enter your character’s name, race, and class. Once I have enough of these, I’ll give them to the neural network and see what happens. Edit: wow, over 3500 responses so far! (Check them out at this link) Keep them coming!
NEW POLL! Neural networks want to hear your character’s backstory! Submit as many as you like. https://goo.gl/forms/ReInNw0Tz0mwzTLO2
From @gobo_tabbycat: “…why do you have to throw my moussey?” #catsofinstagram [source: http://ift.tt/2zJMBCS ]
weird how throughout the history of the western world (as far as i know!), ‘learned’ people, scientists, historians, etc. were usually tied (in varying degrees) to some kind of patron; eg greek slaves in rome, kepler on the court of some king. (s> if they were so smart, why couldn’t they be the patron? (/s>
There are a few reasons. First off, a lot of social structures of the time tended to have anti-meritocratic effects. While being smart was certainly an advantage in keeping crown or title and advancing the interests of one’s nation and dynasty, most people didn’t get into hereditary monarchy or peerage through intelligence.
Another issue, which crops up for intellectuals in the modern day as well, is that a lot of these forms of learning only provided wealth and power indirectly, or with a substantial delay. While basic scientific and mathematical research, stuff like calculus and relativity, has undoubtedly enabled the creation of an enormous amount of value, it required a great deal of practical implementation to do so. And even the more immediately practical inventions and discoveries required substantial resources to create, which in turn had to come from people who already had them.
And of course, while it wasn’t the typical case, there were some intellectuals who did manage to achieve wealth and power through their learning. The classic example is Thales of Miletus, a pre-Socratic Greek philosopher whose knowledge of meteorology allowed him to corner the olive market and become independently wealthy.
“if you enjoyed this proof that the square root of two is not a rational number, why not subscribe to my Patreon! and please remember to like this speech and share it with your friends.”
An interesting case is that of renowned french chemist Antoine Lavoisier, who gave hydrogen, oxygen their names (also the french name of nitrogen, “azote”), and also was one of the discoverers of the law of conservation of mass, which he stated eloquently in a sentence that is still famous to this day in France : “rien ne se perd, rien ne se crée, tout se transforme” (”nothing is lost, nothing is gained, everything is transforming”).
Now the reason for why Lavoisier was able to do all that research was that after studying law, he had been able to become a “fermier général”, which literally translates as “general farmer”, but actually what he was farming was taxes, he was a tax collector for a while and then worked in the fiscal bureaucracy of the kingdom (on the royal monopolies of gunpowder and tobacco).
As such he had access to the highest precision scales available at the time, and to quite a bit of money, so he was able to do a lot of important, groundbreaking fundamental scientific research.
And unfortunately for Lavoisier, the French revolution came.
The hated ferme générale was abolished, and a few years later during the Terror, he was convicted and sent to the guillotine.
A very famous (and probably very apocryphal) story states that when Lavoisier was condemned, he pleaded for a delay in his execution, so that he could finish some scientific work, to which his judge is said to have answered that “the Republic does not need savants” (”la république n’a pas besoin de savant”, another famous sentence).
Thus ended the life of one of the most important scientists of the 18th century, because of his day job as a civil servant in fiscal bureaucracy.
Now that I think about it, he was probably one of the first scientists to be paid by tax revenue, which in his case he personnally collected.
bee-squared-official
