Morning news on July 20, Google's DeepMind co-founder Demis Hassidic recently said at the artificial intelligence conference in New York that Google is using technology from Deep Mind to optimize data center energy consumption. , thereby saving a lot of energy costs. The most well-known, it should be that they developed the artificial intelligence Go application Afghan to defeat the world champion Li Shishi.
DeepMind is an artificial intelligence laboratory based in London, UK. Their research direction is to develop a universal self-learning algorithm. London has about 140 members dedicated to “solving intelligence†and is arguably one of the most interesting technology companies in the UK. . In 2014, the company was acquired by Google for £400 million, which was equivalent to more than $600 million at the time.
In recent months, Google has “find a job†for the DeepMind artificial intelligence system, which allows it to manage some of its data centers by manipulating computer servers and related devices (such as cooling systems) to reduce energy consumption. The technology used in this system is similar to the technology of DeepMind's self-learning Atari video game.
It is reported that Google consumed 4,402,836 megawatt hours of electricity in 2014, equivalent to the total electricity consumption of 366,903 US households a year. A large part of this electricity comes from Google's data center, which supports Google's global coverage of web services and mobile applications.
Google said in a statement that the DeepMind artificial intelligence system increased power usage efficiency (PUE) by 15%. Power usage efficiency measures the power used by Google on data center computers, compared to the power used on supporting infrastructure such as cooling systems.
For Google, saving a few percent of power consumption means huge economic benefits. According to the US Energy Information Administration, the price of commercial electricity in the United States is usually between $25 and $40/MWh. If you can reduce the power consumption of the data center by 10%, then after years of accumulation Google will be able to save hundreds of millions of dollars. In this way, it took a few years for Google to buy it back!
Not only that, artificial intelligence does a lot of things, and then it is the turn of mental health.
Earlier this year, Deep Mind established a health department. Earlier this year, Deepmind UK launched the Health division artificial intelligence medical system.
Deep Mind Health designed a software called Streams. In the pilot project of the Royal Free Hospital (Loyal Free), it allows doctors to quickly view the results of the medical examination. Chris Laing, deputy director of patient safety at the hospital, said Streams allowed him to view the patient's blood test results in seconds, determine the risk of acute kidney damage, and improve patient care.
In July, it was announced that it will cooperate with the Eye Hospital of Mufield to apply artificial intelligence to medical research such as diabetes and retinopathy with the help of the Clinical Medicine Mobile APP.
Although artificial intelligence is still at the primary level in terms of medical assistance, its development potential based on deep learning systems is enormous. In recent years, in addition to traditional statistical analysis algorithms, DeepLearning and artificial intelligence have become powerful weapons for acquiring knowledge from massive data, and have gradually shown advantages in medical research applications based on big data.
Let me give you a few more examples:
IBM - IBM announced in 2015 that its supercomputer Watson began providing genomic data analysis to help doctors provide the basis for the treatment of cancer and other diseases. Compared with the traditional doctor team, Watson can complete artificial gene sequencing analysis and other cancer treatment diagnosis within three minutes, which is more than 30 times higher than the current labor efficiency.
Deep Genomics - A Canadian company, Deep Genomics, uses deep learning techniques to find mutation-related variants from the non-coding segments of the genome. The first product, SPIDEX, simply analyzes the sequencing results and cell types to analyze the effect of a mutation on RNA cleavage and calculate the relationship between the mutation and the disease.
Carbon Cloud Intelligence - led by the former Huada Gene CEO Wang Jun, aims to collect a variety of vital signs and information big data, and then based on this data, to build a kernel model of artificial intelligence, and then It docks and does such an integration. It can even interpret human genetic and health information, alert health risks, conduct precision medicine and personalize medicine.
Atomwise - Want to use supercomputers, AI and sophisticated algorithms to simulate the process of drug development, thereby reducing R&D costs. With the services provided by Atomwise, you can predict which new drugs are really effective and which are not.
The Human Diagnosis Project - is designed to map the symptoms of all human diseases, similar to the "Human Genome" program for disease detection.
Chrono TherapeuTIcs , a smart wearable device called SmartStop, is designed to help smokers quit smoking 24 hours a day when smokers are addicted to cigarettes.
Ekso Bionics - Developed a wearable bionic mechanical leg that uses sensor control to help people regain their walking ability.
Ginger.io - Collect and analyze data such as text messages, calls, locations, etc. with the same smartphone to diagnose patients' mental health problems.
3Scan - an ultra-high resolution 3D tissue imaging platform that creates data for the entire genome sequencing level. The doctor can apply him to detect the disease.
FoundaTIon Medicine - Analyze the patient's genome and tumor-bearing tumors to create personalized treatments for patients.
This year, MedPeer (located in Tokyo) published a questionnaire survey of 3,700 doctors. The results show that about 70% of doctors believe that they will enter the era of artificial intelligence medical treatment within 20 years. Many doctors say that the most important benefit of artificial intelligence is the high rate of diagnosis for intractable diseases that require specialized knowledge. These intelligent solutions, while innovative, lack the trust of patients. The algorithmic output of artificial intelligence should help doctors make more informed decisions rather than directly specifying treatment prescriptions.
Characristics
â—High saturaion magnetic induction that can reduce the volum and weight of the device
â—High permeability and low coercivity that can improve the efficiency and reduce copper loss
â—Low loss that can reduce the transformer temperature
â—Stable temperature that can keep working at -45℃-130℃
Application fields
It can be widely used to the Inverter Welding machine power supply,X-ray, laser, communication power supply,UPS and high frequency induction heating power supply ,charging power supply,electrolytic and electroplate power supply,frequency control of motor speed power supply and plasma cutting machine.
Iron Core For Inverter Welding Machine
Low Loss Magnetic Ring,Versatile Magnetic Core,High Property Iron Core,High Saturation Iron Core
Anyang Kayo Amorphous Technology Co.,Ltd. , https://www.kayoamotech.com