So I have a STUPID question about Machine Learning.

And I am being serious when I say this.

I want to get into machine learning but I really don’t want to accidentally create the AI that kills us all.

I’m not trying to boast my abilities or anything I’m not that great but I just

one don’t completely understand how machine learning works?

And two how do I keep it from learning more than I want it to??

I’m not trying to be stupid I’m just trying to understand so I don’t make anything that I regret.

Hmmmmmmmmmm....

It's just math don't be afraid.

1. current machine learning is like playing around with ant's brain, or even weaker, when you compare the number of neurons. so no worries about skynet.

2. you supply it with training data, it only learns from what it's given, it doesn't proactively go out of its way to learn by consuming random data around.

You should never be afraid of matrix multiplication

Just pull the plug when it happend. Youll be safe.

Machine learning for the layman -
https://devrant.com/rants/2031677/...

Don't worry you have to tell machines what to do. If you do not set open ended tasks machines tend to behave. If you create a machine that has no stop condition it's good practice to add a kill-switch. As long as you do not put your buggy code in unsupervised charge of nuclear weapons we should be fine.

@M1sf3t oml man I’m workin on sumn but I’m just trying to get info before I decide to get into ML.” 😅😂

@Midnigh-shcode alright so it only learns what it’s told to learn. Alright that’s a little relieving

Surely, you must be kiddin'. Machine learning is nothing but a glorified classification math. 3/4 of the time you will be strugling to make it do anything at all, 1/4 of the time it does what you want with 95% accuracy given the task was something ridiculously simple and your learning data perfect. There are no evil AI's and none of them can get out of control. You're worried about a non-existing problem. If you really are serious with this question then I strongly suggest you start learning ML and messing around with it and you'll understand what I mean

@M1sf3t those things weren't created by accident. The algo does exactly one thing. Its not smart, it doesnt know anything. Its essentially a series of if/and gates. Neural nets are not magic

@Nanos that took a turn

@M1sf3t I will notify you when I finish a project 😂

@M1sf3t damn I felt that one

@Bubbles yeah, but i disagree on it being relieving.

think of all the content-filtering neural networks at google and facebook being trained to filter out the worst posts and comments and videos - being trained on the worst the humanity has to offer...

*that* actually smells of skynet.

I personally prefer to call it statistics with extra steps instead of ML
Except when boasting about my projects

@SalsaBoy XD

And now, I'll proceed to go back to my work before rutee07 notices that I'm back :P
Sorry for cheating rutee07
Don't tag her