Other than machine learning, what other technology produce results

>>62103045
I mean stuff like CRISPR is pretty much like that.

>>62103063
How so? Isn't it just an easier way to edit genes?

>>62103074
Yeah but imagine what happens if you figure out how to create a human that can think on a plane hundreds of times higher than others.
Imagine suddenly being able to craft an organism for any situation.

File: 1503465519391m.jpg (42KB, 1024x576px) Image search: [Google]

42KB, 1024x576px

>>62103123
Go to bed, Rick.

>>62103045
ICO's. you make a digital currency and get millions of dollars

>>62103405
xD

>>62103405
ECKS DEE

>>62103063
Idiot brainlet

>>62103045
Nanomachines, son.

>>62103045
A sufficiently long regular expression.

>>62103045
ML isn't really black box magic, neural networks are.

>>62106328
The most relevant, state-of-the-art ML stuff are built on neural networks, deep learning for example.

>>62106342
That's what I said. ML != neural networks. There are random forests, decision trees, support vector machines and other algorithms used in ML, which are not black box magic.

>>62106395
>mfw these summerfags starting threads about ML think K nearest neighbor, beam search, and boosting are "black box magic"

I'd probably throw genetic algorithms in there as well, but nobody cares about those since recurrent and convoluted neural networks are the hotness right now in data science.

>>62103045
Brainlet here. Why is machine learning a hot topic now? Why couldn't they use algorithms instead of predicting an outcome with statistics?

>>62106539
Predicting an outcome with statistic is using an algorithm. A machine learning algorithm. Logistic regression is probably the most simple one, look it up.

>>62106539
Because theres a lot of things where a rule-based algorithm would be too difficult/impossible to make, instead, you just research a general purpose algorithm that reads training data, "learns" it structure and patterns and apply it to new data, without the need for you to code the pattern detection yourself.

Example: You want to make an algorithm that tells you if an image is a cat or a dog. It would be very complex to make a conventional rule-based algorithm for that, since you would have to deal with pixel-by-pixel shit. Instead, with machine learning you just feed a lot of images correctly labeled cat or dog and the algorithm figures out by itself how to classify new, unlabeled images.

Thats why ML is getting huge nowadays, with giant datasets and computational power we can make a lot of shit that is just not viable coding manually.

Take a look at this https://www.youtube.com/watch?v=t169yNXX4oU

>>62103123
Thought you said 'craft an orgasm for any situation' and I was all in.

File: 1491221061481.jpg (208KB, 1000x1000px) Image search: [Google]

208KB, 1000x1000px

>>62106539
We finally hit the point in computing that you can brute force statistical maps for input/output fields w/o city sized computing equipment. It's all the rage because its a convergent algorithm whose brute forcing scales.

> Brainlets finally found a way to mimic intelligence without having to do the hard work required to understand it.

Caveat being when you expose such systems to data sets and output requirements they've never seen. Then the whole thing falls apart which is why it was later classified as weak/brittle AI.

None of this is black box magic as it just straight forward brute force statistics. The real black magic is coming.

>>62104701
>considered black box magic?
>considered
Please leave and never come back.

>>62106478
I'm still amazed at how well simanneal performs even with decently complicated "step" and "evaluation" procedures
I remember using it for finding out the most distant chromapoint from a given one in CIELab with CIE94 ΔE, mostly because color distance is a fuckhuge function and nobody wants to treat it analytically

>>62106692
>with machine learning you just feed a lot of images correctly labeled cat or dog and the algorithm figures out by itself
That explains 4chan captcha

>>62106767
4chan captcha is from google, and machine learning explains 99% of google actually.