What Is DeepMind

DeepMind is a division of Alphabet, Inc.Responsible for developing general-purpose artificial intelligence (AGI) techniques.That technique is also known as Google DeepMind.

DeepMind usage raw pixel data as input and learns from experience. AI uses intensive learning on a neutral neural network, in which there is a model-free reinforcement learning technique called Q-learning.

While the idea of ​​general purpose AI is controversial, Google has determined to install and improve its AI property on a variety of grounds. DeepMind technology has been challenged to learn the game on its own. For example, when the work was done to beat the Atari Games library, he learned to understand the games without changing the code. After a time, AI could play games better and more efficiently than humans.

In pursuing AI borders, DeepMind worked with the defeat of the Board Game Go. Go is a computational challenge for AI, the main reason is the complexity of the selection among the vast number of possible steps in the game. DeepMind developed a special project called Alpha, a computer program designed to play board games. After several versions of learning under the supervision of the AI ​​model, AlphaGo gave the best number 1 player to the world in 2017. In the same year, DeepMind released a new version called Alpha-Zero, in which unproductive re-enforcement learning was used, which continued to be superior to previous versions of the program.

Outside of the game, DeepMind was used to improve the efficiency of the previously optimized data centers on Google. DeepMind was able to improve the efforts of previous experts by 15%, thereby reducing the cooling cost by 40%. It is also used to develop a Google Assistant and helps in making personalized app recommendations in Google Play.