Is it possible to train a neural network to determine the race of a music artist based on its name?

Is it possible to train a neural network to determine the race of a music artist based on its name?

I made a software that removes black peoples music from my library, but new "artists" are popping up like shit in a kid pool, and I am bored to add them to the list manually every time.
I have a few thousand artist to train the nn.

Attached: 1535538365947.png (437x369, 252K)

Just do a google lookup you fucking retard
>HURR DURR I JUST LEARN ML SO MUST NOW USE IT ON EVERY FUCKING THING

good luck when you get to shit like Tracy Chapman

Attached: yep.jpg (684x546, 76K)

>why automate if you can do it manually
Fuck offy stone age needs you

or just look up the name....

looking up doesn't mean doing it manually you peanut brain

If it can remove all of the new bullshits like xtentacle or whatever, TyroneWeed then I am happy with the results

So if it's an unknown artist it searches the internet for it and I decide if I want to remove it . Could work, easy to implement, but still not as good as complete automatisation

Remove yourself from existence.
It'll be easier that way.

t. NPC

so then sounds like you have zero need for a neural network. Just use the metadata tags on songs to remove anything R&B, Hip Hop, Rap, Gospel, etc, from after 1950 or something. Honestly any time anybody says "do we need NN for this" the answer is "nope"

Attached: 1542038517572.jpg (324x451, 14K)

>Is it possible to train a neural network to determine the race of a music artist based on its name?

no it's not.

Yeah, it already does that, but not every mp3 file is completely tagged up(I torrent a bunch of mp3 libraries and let the software remove the shit, or already downloaded songs).
Some might not even have tags

use something like last.fm, shazam, etc to query metadata information for untagged songs.

To be honest I never used NN-s, because it was easier to create a working model, but I wanted to try out because it's used everywhere

It would be easier to create a database of white rappers, filter out R&B/Hip-Hop/Rap altogether (Even pirated music contains this metadeta), and then make exceptions for the white rappers

Mmmm personally go for a webscraper on Google images for all the artist name.

They identify with a Neural network which race is on the picture.

The label the picture

Then use these labeling for automatical deletion

I don't like white rappers either. It's the same retarded noise for me. As a not native English speaker I can switch on or off whether or not I want to understand English (songs).

If you want to use ML or NN, it would be much more applicable to use ML/NN's to recognize voice patterns and linguistics to determine race

Then just put a filter to not listen to rap and create a database of popular artists in the genres you listen too that are black and filter based on that?

Slightly better results for way harder work. Don't think it's worth it

ML/NN is hard work. It's an infant field in CS rn.

Don't think this is going to be easy like other memes in CS.

Yeah I already do that, but I download a bunch of new, unkown(for me) artists to find new and good songs.
I don't really want to listen to the same song more than 6 times in my life