In chess,AlphaZero outperformed Stockfish after just 4 hours

[Henk]

Is there already an AI program generating pop music. For instance did it create a nr 1 hit.

Lately I found a video where someone was trying to make a neural network play bach music. But task wasn't that easy. Sometimes it played fantastic but most of the time horrible. And before that stage was reached it sounded jazzy.

[Ovyron]

Is there already an AI program generating pop music. For instance did it create a nr 1 hit.

Not yet, but unless you're familiar with Bach's work, there's an AI that generates new music in the style of Bach, that you may not be able to tell between a real piece and a fake one.

[kranium]

I don't care that SF lost, it's totally irrelevant in the light of the huge claim by the Deepmind company, the alleged 4 hours self-play, quoting the document again: without any additional domain knowledge except the rules of the game.

Have you already let it sink in what is stated here?

No mobility, no king safety, no passed pawn evaluation, no castling knowledge, not even piece values?

How would that first self-play game look like? Something 1.a3 a6 2.a4 a5 3. b3 b6 etc and how would that lead to anything for the second self-pay game?

And so I voted for option 3.

Monte Carlo search does not use a tradition eval as we know it, so mobility, king safety etc. are irrelevant.

It uses a struct to hold info likes wins, losses, draws, win %, etc.,
then simply references accumulated data for the current position to select the move with the highest probability of winning.

Ivanhoe has a Montecarlo search implementation (with which I'm fairly familiar) and it works quite well.

The default implementation uses a sort of 'searchmoves' algorithm:
go montecarlo cpus 8 min -25 max 325 length 40 depth 10 moves c2c4 d2d4 e2e4 g1f3

Years ago I experimented with a version that would obtain the root move list from current position and actually play a strong game.
If you send it all 20 possible moves from the traditional start position, you'd be amazed how quickly the potential move choices are narrowed down...and it usually plays 1. c4 or 1. e4
I still have it if anyone interested (but it does crash once in awhile).

[Henk]

Ok then somebody else has already succeeded.

[kranium]

The training phase... didn't it consist of 44 million games or something like that? If that's the case, I don't see how they could be played in just four hours.

Like MIlos said:

"4h on 5000TPUs where each TPU is equivalent to roughly 2 new GV100 or 10 1080Ti which is currently the top of the range graphics card normal individuals can afford. So those 4h of training time is like over 30 years of training on 1080Ti."

This is an enormous resource...self-play usually involves lightning games, sometimes as fast as 1 sec + 1 ms inc.
Just do the math and one can see see how it's possible.

[Henk]

One main problem is getting enough training examples. If you need a great many training examples but you can't create them automatically then it almost gets impossible to solve the problem.

[Vinvin]

...
How would that first self-play game look like? Something 1.a3 a6 2.a4 a5 3. b3 b6 etc and how would that lead to anything for the second self-pay game?
...

Sure firsts games are pretty random then come good results with good setup then keeping this setup and then finding even better setup because of better results and loop this a lot of times ...

Hint, have you considered why there is no 8-man TB yet?

If I understand well what you mean : "Because they are too big ?"
TBs store exact values but Neural Networks store "shapes" (they sometimes work great and sometimes work badly).

[yurikvelo]

Please clarify on A0.

Can it analyze arbitrary FEN position or it's learn-tree is based on games of strong engines?

[Rebel]

No mobility, no king safety, no passed pawn evaluation, no castling knowledge, not even piece values?

Yup, I think true Artificial Intelligence has finally arrived, and it can do things like this and others that I would have never imagined to be possible.

Some examples of similar AIs:

AI can extract the style of a photo and turn another photo into that style
AI can learn how to make paintings of any artist of history and use any image to show how that artist would have painted it.
AI takes text as input and creates new photo realistic images indistingishable from actual photos.
AI learns how humans lips move when talking, so it can sync a video of anybody to any audio talking.
AI learns how celebrities look like and can invent new faces for fake ones that look real.
AI learns how art looks like, so it can turn your doodles into works of art.
AI learns how video works, so it can predict the future and create videos from still images
AI learns how images become pixelated when you scale them down and manages to reverse the process, turning pixelated messes into High Resolution images.
AI learns how visual expressions work and can swap the expressions of two people.
AI can turn your sketches into photo realistic images.
AI learns how to play non-deterministic video games just like humans.

All driven by domain specific knowledge, thus off-topic.

[Rebel]

...
How would that first self-play game look like? Something 1.a3 a6 2.a4 a5 3. b3 b6 etc and how would that lead to anything for the second self-pay game?
...

Sure firsts games are pretty random then come good results with good setup then keeping this setup and then finding even better setup because of better results and loop this a lot of times ...

Hint, have you considered why there is no 8-man TB yet?

If I understand well what you mean : "Because they are too big ?"
TBs store exact values but Neural Networks store "shapes" (they sometimes work great and sometimes work badly).

Yep, size.

https://en.wikipedia.org/wiki/Shannon_number

https://en.wikipedia.org/wiki/Solving_chess