Simple Stats Administration: Make Another Proboscis
Janelle Shane from aiweirdness.com is at it again, showing us what the true purpose of a neural network actually is. Previously her research brought us Horble Grey, Turdly and a bevy of other new colours of paint to choose from, as well as metal bands, horror movies, beer flavours and so much more. The Inquirer spotted a new batch of absurdity today, and after the disturbing week we’ve had as far as the security of our computers go, it is very welcome on this Friday.
She fed 190,000 Change.org petitions into a 117M-GPT-2 neural net, built by OpenAI and trained it with GPUs on Google Colaboratory, we can now see what our AI overlords are truly worried about. For instance, we can all get behind the movement for “Taco Bell: Offer hot wings and non-perfumed water for all custom” though some may not want to “Help Bring Climate Change to the Philippines!” Follow the link from The Inquirer to read more, and to get some of the highlights from her previous works.
Janelle has fed 190,000 Change.org petitions into a neural network and set it about making its own. She explains: "The neural net I ended up using is 117M-GPT-2, by OpenAI, which is better at stringing together readable sentences than some of the other neural nets I've used.