Google’s DeepMind co-founder Demis Hassabis (Demis Hassabis) recently said at the New York AI conference that Google is using DeepMind technology to optimize the energy consumption of data centers, thereby saving a lot of energy costs.
in recent months, Google is using DeepMind’s AI system to control some parts of data center, so as to optimize the energy consumption of servers and refrigeration equipment. Previously, DeepMind’s technology was used to teach computers to play Atari games.
said that Kazakhstan Biscay, Google data center technology makes the energy consumption declined by several percentage points. “From the cost point of view, this brings great savings and is also beneficial to environmental protection.”
Google says this optimization has brought about 15% of the energy use efficiency. The energy efficiency index compares the power consumption of computer power consumption with the infrastructure of the refrigeration system.
Google said the average electricity consumption in 2014 was 4402836 megawatts, equivalent to the average annual energy consumption of 366903 households in the United States. A large portion of these power consumption comes from the data center, which is the basis of Google’s network services and mobiles applications.
will reduce the power consumption by a few percentage points for Google’s financial help. According to the US energy information administration, the electricity price in the United States is usually $25 to $40 a megabyte. So, if the energy consumption of the data center is reduced by 10%, then Google can save hundreds of millions of dollars in a few years. Google bought DeepMind at $400 million in 2014, at a time of $more than 600 million.
Google has also used machine learning technology in the data center before. In 2014, Google said it was using the neural network system to predict the changes in energy consumption over time, so that equipment could be more effectively arranged.
The work of
DeepMind is further based on this. The company’s software can regulate the operating mode of the equipment in the data center and improve the efficiency of energy consumption. Kazakhstan Biscay said: “the control of about 120 variables in the data center, including fan, refrigeration system, and windows and other equipment.”
at the same time, says it’s just the beginning of the project. At present, DeepMind knows this method is effective and begins to know what information its AI systems are missing. So, DeepMind may ask Google to deploy new sensors in the data center to help the software get more information.