neural network - AI Weirdness

Pokemon generated by neural network

By Janelle Shane On July 22, 2016 - 1 min read

I’ve been playing around with char-rnn, an open-source torch add-on for character-based neural networks by Andrej Karpathy [http://t.umblr.com/redirect?z=https%3A%2F%2Fgithub.com%2Fkarpathy%2Fchar-rnn&t=ZTNhNTMxMDJmZjFkMDJiMjI1MDZmZjk2ZGUyN2MxZjhiNTk2MWFjMCw0T1VwUVFXVQ%3D%3D] , using it to generate everything from cookbook recipes [http://lewisandquark.tumblr.com/tagged/recipes]

Cooking with Cthulhu

By Janelle Shane On March 19, 2016 - 1 min read

Here’s what you get when you give incomplete cookbook recipes to a neural network trained on the complete works of H. P. Lovecraft: > Bake at 350 degrees for 30 to 32 minutes. Test corners to see if done, as center will seem like the next horror of Second

The neural network doesn’t understand proper nouns.

By Janelle Shane On March 18, 2016 - 1 min read

As the neural network begins to get better at generating cookbook recipes [http://lewisandquark.tumblr.com/post/140219420017/the-silicon-gourmet-training-a-neural-network-to] , it continues to have trouble with recipe sources - short and highly varied, they’re a challenge for an algorithm that learns by repetition. Still, it does its best: Source: Carrots

Disturbingly vague ingredients generated by neural network

By Janelle Shane On March 15, 2016 - 1 min read

This neural network, a learning algorithm trained on 30MB of cookbook recipes [http://lewisandquark.tumblr.com/post/140219420017/the-silicon-gourmet-training-a-neural-network-to] , generates new recipes based on probabilities. The resulting ingredients, while their words are individually probable, can end up disturbingly vague. “Yeah… I’m pretty sure this recipe’s gonna contain some…

The neural network invents fancy new ingredients

By Janelle Shane On March 12, 2016 - 1 min read

I’m training a neural network to generate recipes [http://lewisandquark.tumblr.com/post/140219420017/the-silicon-gourmet-training-a-neural-network-to] based on a database of about 30,000 examples, and although it sometimes comes up with ingredients that sound like a bad idea, sometimes it invents some that you could plausibly ask for at

Superheroes designed by neural network

By Janelle Shane On March 10, 2016 - 1 min read

I trained the neural network to generate superhero names, based on the list from this site. [http://www.superherodb.com/characters/] I thought the database was going to be way too small, but the network proved me wrong. Speet Stank Red Fart Mister Man Rad Food Sapgirl Woop Ann Man

The neural network has bad ingredient ideas

By Janelle Shane On March 08, 2016 - 1 min read

I’m training a neural network to generate recipes [http://lewisandquark.tumblr.com/post/140219420017/the-silicon-gourmet-training-a-neural-network-to] based on a database of about 30,000 examples, and one great (not great?) thing about it is it comes up with new ingredients that I’m pretty sure aren’t in the list:

The neural network gives bad cooking advice

By Janelle Shane On March 04, 2016 - 1 min read

I’ve trained the neural network on about 30MB of cookbook recipes [http://lewisandquark.tumblr.com/post/140219420017/the-silicon-gourmet-training-a-neural-network-to] , but its cooking advice still doesn’t seem that great. I’m not sure I should do or add any of these things. > 1 cup cherry seeds 42 cup milk

Tag: neural network

Pokemon generated by neural network

Cooking with Cthulhu

The neural network doesn’t understand proper nouns.

Disturbingly vague ingredients generated by neural network

The neural network invents fancy new ingredients

Superheroes designed by neural network

The neural network has bad ingredient ideas

The neural network gives bad cooking advice

Writing

Subscribe

Recent Posts

Botober 2024

Bonus: "Ignore all previous instructions" gets weirder

An exercise in frustration

Bonus: A unicorn goes downhill

Hidden 3D Pictures

Hidden sheep

Follow

Botober 2024

Bonus: "Ignore all previous instructions" gets weirder

An exercise in frustration