[Art]

...or do they?
According to Google, they now have it.


https://futurism.com/google-has-started-adding-imagination-to-its-deepmind-ai/

[ivan.moony]

The essence shouldn't be too hard. Neural networks recognize from examplars to population. All they had to do is to reverse the process to imagine from population to examplars recursively (they have to randomly, or some other way pick each one of examplars in a pyramid chain). Though I don't know how much it complicates NN algorithm that is a complication record keeper in its own way in back-propagating learning process.

[selimchehimi]

It doesn't surprise me that Google announced that!
As far as I know, Deep Mind (acquired by Google in 2014) wants to create a kind of Artificial Hippocampus which is responsible for memory, attention, imagination and planning.
I've made a video about what Deep Mind is planning to do to solve AI, you can check it here :

[Art]

Selim,

Your video was nice. I especially enjoy your site's opening graphics. Very well done!

[selimchehimi]

Hey, thank you so much, I really appreciate it! And I'm very glad that you liked

[ivan.moony]

It was incredible how Deep Mind concluded that digging a tunnel in bricks in the Arkanoid game is the best tactics to make a progress. I'm seriously stunned.

[Edit] I wonder what NN theory is behind that conclusion.

[selimchehimi]

It was incredible how Deep Mind concluded that digging a tunnel in bricks in the Arkanoid game is the best tactics to make a progress. I'm seriously stunned.

[Edit] I wonder what NN theory is behind that conclusion.

I don't understand what you mean, that it's not the right conclusion?

[ivan.moony]

It was incredible how Deep Mind concluded that digging a tunnel in bricks in the Arkanoid game is the best tactics to make a progress. I'm seriously stunned.

[Edit] I wonder what NN theory is behind that conclusion.

I don't understand what you mean, that it's not the right conclusion?

I think it is a cool conclusion. In deed, digging a tunnel is a fastest way to get to the next level.

[selimchehimi]

Oh ok! Moreover, the researchers at Deep Mind are now playing the game better thanks to the AI