The development of Microsoft machine learning begins with a polygraph

Recently, Microsoft's Jennifer Marsman tested his own lie detector on his own boss.

“Do you think your company is the best company in the world?”

"of course"

"Ha! According to the response of the polygraph, this is a bit of a dim."

"I will promote me this year?"

"of course"

"This doesn't sound like a fake."

However, Microsoft is not playing law enforcement games. The 37-year-old Marsman is a "principal developer evangelist" whose job is to spare no effort to promote machine learning - a form of artificial intelligence that uses data to predict everything from quarterly sales to when cows are pregnant. and many more. (Translator's Note: Technical evangelists are the most front-line and most important "translators" who can interpret technology in an easy-to-understand way to people from different fields to gain their support for products or technology. People who know the technology and can dig out the stories behind the technology, they can stimulate people's passion for a product.)

The singer designed by Marsman combines the algorithm with 14 head-mounted devices that are used to monitor the brain, which is like a little trick at the party. Marsman uses this to show software developers how to use the Microsoft Azure Machine Learning tool. Marsman plays an important role in Microsoft. Microsoft started early in the machine learning arena, but now faces competition from Google and Amazon Machine Learning Commercialization.

This risk is quite high. In the next few years, machine learning will change the world – it will increase the level of computer intelligence exponentially, and help cut company costs, predict what is worth investing, and what is worth investing heavily. Anurag Rana, an analyst at Bloomberg Intelligence, called the technology "the most important factor in distinguishing software companies." After leaving the machine, he said, "You can't sell the product."

Although Microsoft has been working in machine learning for at least 20 years, departments such as Office and Windows have used their predictive features with care. “Many people responded by 'We know how to do it, why do you still have to use data to question my point?'” Pedro Domingos, a professor of computer science at the University of Washington, said that he wrote a book on machine learning – Algorithm The Master Algorithm.

When Microsoft tried to catch up with Google with Bing search, Microsoft really embraced machine learning technology. Two years ago, Satya Nadella served as Microsoft CEO. Prior to that, he set the engineering and technology strategy for the search department and sprinkled machine learning technology into the company's products. "Machine learning is deeply embedded in Microsoft, and Microsoft is now in this position," Domingos said. “They are investing heavily in machine learning to make the field less ridiculous.”

Just as Google and Apple use machine learning to improve their products, Microsoft has integrated this technology into its operations. This is not just about cost savings, it helps the company to operate better; the more Microsoft uses this technology, the easier it is to explain and sell to customers. “Customers are confused,” said Joseph Sirosh, who in 2013 hired him from Amazon to take charge of Microsoft machine learning projects. “It’s challenging to move forward in a piece of incomprehensibility. At the same time, the internals are also difficult. Sales people have to convince consumers and explain all the usage scenarios to them.”

The finance department of Microsoft's chief financial officer Amy Hood has begun to rely on algorithms—using algorithms to predict sales data and predict the number of licenses in a given time period. “The results are very, very accurate,” says Sirosh. “Amy Hood is a great fan of machine learning. She knows that the machine learning model predicts quarterly data and she can sleep well.”

Microsoft also uses algorithms to predict how many servers will need to be purchased as the data center expands rapidly, and help salespeople identify key customers. Sirosh said that even the old products, such as a financial software acquired in 2002, have been blessed by machine learning. Microsoft's Cotrana AnalyTIcs Suite allows users to create such a tool by themselves.

25-year-old Ram Shankar Siva Kumar claims to be a data cowboy and a member of the Azure Security Data Science team. He uses machine learning algorithms to predict suspicious behavior in Microsoft networks. Once you know what you need to find, Microsoft's security team can quickly find the source of the attack, and Kumar must find them before anyone knows about it.

In order to train the algorithm's ability to identify harmful behaviors, he paid to encourage Microsoft's Red Team hackers to attack the network and dangerous reports from the Microsoft Security Center to train the algorithm with actual attacks. This helps him build a model to identify real vulnerabilities.

Microsoft's technology is being used in all walks of life. Japanese farmers use technology to track cows, and cows are more likely to move around when they are pregnant, so they can conceive cows at the best time. An Australian winery also uses a similar algorithm to predict grape yield. A hospital about an hour's drive from Microsoft uses Azure tools to find out which heart patients are more likely to need a hospitalization exam. The Norwegian eSmart system uses Azure machine learning to predict energy grid usage and turn off home heating when demand is high.

Matt McIlwain, general manager of the Madrona Venture Group in Seattle, believes that Microsoft's machine learning technology is no worse or better than its competitors' technology. But he also believes that Microsoft still lacks brand awareness and is currently catching up. “How do people discover that Microsoft really has great machine learning skills?” he said. "Microsoft must sell itself out."

This is the entry point for Jennifer Marsman. She traversed the world, demonstrated her lie detector and promoted the potential use of machine learning. There are many discussions on medical applications. People have been asked to use this technique to predict epilepsy, monitor elderly people who depend on assisted living facilities and decide whether the athletes injured in the game go directly to the hospital or return to the field to continue the game. “I have the coolest job at the company.”

Bi-Concave Lens

The double-concave lens is similar to the plano-concave lens. The focal length is negative, but the parallel incident light diverges outward. The two side surfaces of the double-concave lens have the same radius of curvature, and are generally used for beam expansion and projection.
The image formed by the biconcave lens is always smaller than the upright virtual image of the object. The biconcave lens is mainly used to correct myopia. Myopia is mainly due to the deformation of the lens, which causes the light to gather in front of the retina prematurely. The double-concave lens plays a role of diverging light. The double-concave lens becomes an upright, reduced virtual image, which makes the image distance longer and just falls on the retina.

Double Concave Lenses,N-Bk7 Bi-Concave Lens,Quartz Biconcave Lens,Optical Bi-Concave Lens

Bohr Optics Co.,Ltd , https://www.bohr-optics.com