How ‘Microsoft Flight Simulator’ became the ‘Game of Life’ with Azure AI

After some investigation, Newman realized that the data set on the Bing map covered the entire planet. The only problem? It was in 2D. After using some of that data to create a flying 3D version of Seattle, Newman returned to the Azure team to design a machine learning system to transform the entire planet into a giant 3D model.

“AI has grown tremendously over the last few years,” Eric Boyd, CVP of Azure AI, said in an interview. “It is actually driven by the massive amount of data now available, combined with the massive calculations that are in the cloud … the results you can see are so beautiful that you can now come up with the means to identify individual trees, grass and water for every square kilometer of the planet, and then 3D Use it to create models. “

The integration of the Assyrians is beyond the shape of the world. It also powers aircraft control voices using AI speech generation technology, which is almost indistinguishable from humans. Many players may think that Microsoft relies solely on voice actors.

Boyd says machine learning capabilities have also developed dramatically since the company began exploring ways to bring Azure AI into the game in 2016. “The AI ​​algorithm space has really grown over the last several years,” he said. “So visual algorithms are widely used to identify these different trees and buildings and to classify them correctly, and they have come a long way.”

Because it leans heavily on the cloud, Flight Simulator A “game of life” in the true sense, Newman said. All of the machine learning methods that the company relies on will improve steadily over time as the company solves errors and upgrades the machine. (Also be aware of potential problems such as spelling Built the 212-story tower In Melbourne.) But he points out that the algorithms can be just as good as the source data, so Microsoft is working hard to clean it up as well.