just a version of my project where i test stuff
I think I have a way to make this more accurate im going to try rn the higher the learning rate is, the easier it is for it to comprehend the bigger numbers which i believe is more realistic, because to be able to comprehend bigger things, youre gonna need to have the learning capabilities to do so. But if the learning rate is too high, then it can overshoot or undershoot the estimation. Which shows that if you learn too much, you might dig too deep into things that arent meant to be. this displays that there is a perfect amount of learning that someone or in this case something needs to be able to comprehend something. Not too much, not too little. So basically, this is just a really stupid AI model. maybe in the future i can use something like unity or some game engine to turn this into a sentient robot but i feel like that would be a little hard to do in scratch... okay i js came back to thiws its been like a month and a half since i made it and shared it. i really dont know what this could be used for but i think it could be built upon. i might see what happens if i add more nodes. maybe ill try to make it sentient one day lol how this works: the input is the node on the far left. The nodes in the middle just basically try to comprehend the input, and the last node is the output, basically what it comprehended. The variable learning rate, and RPT can help it be more accurate. TrainedOutput is the new output after it was trained by the learning rate and rpt so it should be a bit more accurate. TO error is just how far off the trained output is.
