Basically all computer random functions are pseudo-random. It’s a question whether something truly random exists, even in life
True.
Thats deep
Update: IT IS FULLY WORKING!!! Just going to add some artsy details and I’ll send it out later today. Tysm everyone for the support!!
Bro… that is really, REALLY polarised thinking. Lol
Cant wait to see it happen!
AI FASHION FROGS #1 - COLOR GUESSING IS HERE!
Help your fashionable froggy friends determine your personalized favorite-color-predicting AI! Answer any amount of guessing questions you want, and see how well the Neural Network can figure you out.
How do I play?
Simply press A to continue dialogue, type in the amount of guesses you want, and select the frogs with the best answers.
How does this work?
Neural networks are a form of Artificial Intelligence (AI) based off of the human brain (neural for the human neurons!) My specific Neural Network is a web of perceptrons with 3 inputs (R, G, B Color), 2 hidden layers with 4 perceptrons each, and one output that guesses from 0-1 how much you like a color. By picking the most correct Networks, those two are mixed together to form new networks that should be even more accurate, until you have a completely correct network. To ensure the user doesn’t get trapped with the same 5 frogs with the same bad opinions, the net has mutations that change about 5% of the net each time. (I’m not an expert, go ahead and search up ‘Neural Networks’ or ‘Perceptrons’ to learn more!)
What is in the future for AI Fashion Frogs?
Well I’m so glad you asked- er, clicked. Your treetop amphibian friends aren’t done helping you with your style! I’m hoping to make more AI games with them. The next that is currently in planning will demonstrate the frogs designing you entire outfits- froggy tailors! No promises, though. I would love suggestions!
yayyy!
i rarely saw any frogs saying love.
How did they match all of my fav colors?! That’s soooo cool!
From my IRL playtesters I get a lot of feedback like this- “Not enough frogs are saying ___!”
The AI is the one choosing the answer, so, unlike my sister has said, I cannot just break into the code and make them say different things. You will find that the more time you spend training the frogs, the more times the right answers will come up!
You could also have been picking the frogs that are saying ‘Ribbit?’ which is my lazy way of saying ‘Don’t pick this frog- it has mutated so far that it has extremely hateful or extremely positive opinions!’
I actually do have a solution for this- by not showing the user the frog’s answers and instead having the code pick the two most accurate frogs, the user doesn’t get frustrated at the training process!
YEAH!! Can you BELEIVE that only 41 neurons can distinguish that?? I think it super awesome. I can’t wait to see them go farther- Whole outfits, perhaps!
Ah, ribbit does that apparently I guess? But then I have no other choices.
Yeah. Sounds like you got stuck in a loop there : (
I’m really glad for the feedback though- this will be super helpful for my next AI game!! /gen
Alright!
Just to know, since you are turning the AI into an extension are you going to continue the AI speech?
No, as the speech wasn’t AI and sarge is actually doing the same but much, much better! (Not a self-depricating thing, I mean he has the technical knowledge and tools X3)
When should he be done with the AI Speech?
Hmm, I don’t know. He is in UAT2 right now so probably after that