Artificial intelligence DeepMind no longer needs people

Artificial intelligence DeepMind, developed by Google specialists, now does not need the support of its creators in order to develop and improve further. This was achieved with the introduction of a new system Differential Neural Computer (DNC), which combines the ability of computers to store large amounts of information, logical skills of artificial intelligence, as well as the ability of a neural network to quickly search the data store for the necessary fragments."This model is able to learn by examples as a neural network, but at the same time it can store large amounts of data like computers," explains the leading developers of DeepMind Alexander Graves and Greg Wayne.Like the brain, a neural network uses a network of interconnected nodes in order to stimulate the correct centers necessary to accomplish the task. Artificial intelligence optimizes the work of nodes in order to find the fastest solution, which in this case will be correct. Over time, the AI ​​will use the experience it has gained in order to find the right answer faster and more efficiently.


Developers from Google gave a clear example of the work of the updated system. After the DNC explained how the genealogical tree is arranged, artificial intelligence has managed to build multiple connections between all its participants, while this is processing  it`s optimizing and ordering their memory, so that in the future will help to find necessary information as soon as possible . After this, artificial intelligence was shown a map of the London subway, and he independently began to look for additional ways and connections between them, in which case it was possible to promptly route the right route from one point to another.

Instead of calculating every possible outcome of the situation in search of the right solution, DeepMind can find the answer, based on its previous experience, discover it in its internal memory and link it to the current situation. In this case, people do not have to give him hints or feed new input data in huge quantities. For scientists, this is a huge achievement in the field of artificial intelligence development. This is how DeepMind worked, when earlier this year it defeated the world champion in Go. But this is just the beginning. And what will happen next - time will tell.