Yes the DeepMind of Google are now able to teach themselves based on the information they already have.
In an important step in artificial intelligence, the Alphabet hybrid system, called the Differential Neural Computer (DNC), uses the data already stored on conventional computers and corresponds to a very clever AI and a neural network capable of fast parsing.
"These models can learn from examples, like neural networks, but they can also store complex data, just like computers," said the researchers by DeepMind Alexander Graves and Greg Wayne.
What does this mean;
Just like a brain, the nervous network uses interconnected nodes to stimulate specific centers needed to complete a task. In this case, AI optimizes nodes to find the fastest solution and deliver the desired result.
According to the researchers, over time, it will use the data obtained to make it more effective in finding the right answer.
So instead of having to learn every possible outcome to discover a solution, from the information the developers feed it, DeepMind will be able to find answers from previous experiences of him, from his inner self μνήμη and not from external programming.
Depending on the viewpoint everyone sees, the event could be a serious turn to the smartest AI that one day they can think and learn, just as people do.
https://deepmind.com/blog/differentiable-neural-computers/