Janelle Shane @JanelleCShane I blog at aiweirdness.com/. My book "You Look Like a Thing and I Love You" is out Nov 5th! Research Scientist in optics. she/her. wandering.shop/@janellecshane Oct. 14, 2019 1 min read

Need a Halloween costume idea? Here are some ideas generated by a neural net
 https://aiweirdness.com/post/188342947482/halloween-costumes-by-the-neural-net-gpt-2 

In 2017 I tried to generate Halloween costumes with a neural net that didn't have any prior knowledge of English words. Its costumes often made no damn sense.
 https://aiweirdness.com/post/166814009412/a-neural-network-designs-halloween-costumes 

In 2018 I tried generating Halloween costumes with a slightly bigger dataset, but the neural net still couldn't do more than remix letter combinations.

But @susanthesquark of @nytimes arranged for a Professional Illustrator & Jessia Ma nailed it:
 https://www.nytimes.com/interactive/2018/10/26/opinion/halloween-spooky-costumes-machine-learning-generator.html 

My 2019 Halloween costume attempt uses the same dataset, but a neural net (GPT-2-355M) that has picked up a lot of context from the internet.

This is most obvious here. It grabbed words like bitcoin & titmouse that weren't in the example costumes. Compare to 2017 (conchpaper??)

When I increase neural net sampling temperature, the output conforms less closely to the training data.

In 2017/2018 that meant more nonsense words.

The 2019 neural net reverts to its internet training data, introducing more Game of Thrones, Frozen, and racism. Oh, internet.

It is super interesting to see the form chaos takes in a small vs huge neural net.

Unfiltered text from 2017 char-rnn vs 2019 GPT-2, both trained on the same dataset of halloween costumes & sampled at chaotic 1.4 temperature.

reverting to nonsense vs to celebrity references

i have been looking for an excuse to post this and it's friday night so


You can follow @JanelleCShane.



Bookmark

____
Tip: mention @threader_app on a Twitter thread with the keyword “compile” to get a link to it.

Enjoy Threader? Sign up.

Threader is an independent project created by only two developers. The site gets 500,000+ visits a month and our iOS Twitter client was featured as an App of the Day by Apple. Running this space is expensive and time consuming. If you find Threader useful, please consider supporting us to make it a sustainable project.