Is Machine Learning a meme?

Is Machine Learning a meme?

Other urls found in this thread:

courses.csail.mit.edu/18.337/2015/docs/50YearsDataScience.pdf
washingtonpost.com/news/innovations/wp/2017/09/25/teenage-suicide-is-extremely-difficult-to-predict-thats-why-some-experts-are-turning-to-machines-for-help/?utm_term=.2ef229571f79
twitter.com/NSFWRedditVideo

No

yes, even ML techniques are retreating towards axiomatic methods (capsule learning)

Yes. cf. Dawkins **THE SELFISH GENE**

I will make a machine make causal inferences I swear there is a way I will show you all!

Image posted is a neuroscientist, just so everyone knows.

yeah its a meme. Computers are far too weak right now. Wait for quantum computers, and then lets talk. Unless of course, all you want is some shitty scripts and not something awesome like simulating human consciousness.

Post Grimes

Uh, computational theory disagrees.

Yeah, good luck building a city filled with super computers and nuclear reactors. A good quantum computer would simulate human consciousness and fit in your desktop

remember to be nice to Grimes

yes.

courses.csail.mit.edu/18.337/2015/docs/50YearsDataScience.pdf

stfu retard

You have no idea what you're talking about.

Quantum computing is no where near that level yet. The main hurdle is the engineering side of things, which may not advance for decades, or centuries.

Every time I see her I imagine her having a really hairy, sweaty pussy.

And I like it.

Lrn2meme fgt pls

The main difficulty is actually error correction and the programming side. Look at the qubits that Google and UCSB have manufactured. None of the difficulties they're experiencing are engineering or physics related.

>Computers are far too weak right now
>Wait for quantum computers
We seriously need that brainlet filtering calculus captcha already.

Hmm, I must be a couple years late or something. I could have swore the main issue was cooling and actually designing the logic systems/qubits. Maybe a career in applied math focused on quantum algorithms wouldn't be pointless after all...

t. Undergrad mathematician who used to be obsessed with Shor and Grover

Are you a meme?

>The main difficulty is actually error correction and the programming side.
>None of the difficulties they're experiencing are engineering or physics related.
Implying computers that keep introducing errors into your system are not a physics or engineering problem.

If we'd be able to build reliable QC, you wouldn't need incredible amounts of Error correction because you wouldn't even have all those errors to start with. Since we can't build QC that well rn, we choose to try and circumvent the issue by adding ECC.

I'd go for it but fucked up with my PhD choice, now it's too late to switch.
Quantum algorithm development will be a hot af topic once QC get useably fast and large (8qubits is the largest I know of which of course is pointless for any real world applications)

I guess trips don't lie... Either way, I'll be watching developments in the field closely.

No, it's a real thing with a misleading name overhyped by journalists and then the "LE SCIENCE xDD BAZOOKA!!" crowd

To know if something is a meme or not you just have to ask yourself: Has it been on mainstream news? Let's investigate that.

Oh, here it is: Washington Post. washingtonpost.com/news/innovations/wp/2017/09/25/teenage-suicide-is-extremely-difficult-to-predict-thats-why-some-experts-are-turning-to-machines-for-help/?utm_term=.2ef229571f79

So yes, it is a meme.

Only field where you can research using tons of mathematics and actually get funding for it and work outside academia.

How can you say its a meme ?
Post one speech recognition technique that isnt ML. Post one computer vision algorithm that is not ML...

>Is [broad topic that covers a variety of techniques which leads to actual unprecedented results some of the time] a meme

Ask Richard Dawkins.

Always

*snnnnniiiifffff*

It is a meme but it is also pretty useful

holy fuck, he explicitly called "big data" a meme

...

Data "Scientists" BTFO

We need a captcha that can use a neural network to decide what general topic a thread is about, then gives you a problem relating to that topic.

>Limited TOC
And this proves...?

>capsules
>axiomatic methods
u wot m8?

No, it's literally the only thing that's going to matter in 10 years

>already used in bussiness with success
>meme

choose one. It's a bit overhyped but it's good shit.

Shoo, shoo, grimefag, back to your containment board you go.

Doesn't everyone know that a Data Scientist is just a statistician who knows how to market himself?

Pals

>the chad data scientist
>the virgin statistician

No they will be our slaves within the next few years.