The American space agency NASA and the technology company Google have identified an eighth planet in a faraway solar system.

    That solar system now has exactly the same number of planets as our own.

    Machines made the surprising discovery, not human researchers.

    NASA and Google representatives made a joint announcement about the discovery on December 14.

    The newly discovered planet orbits the star known as Kepler-90. The system is about 2,545 light-years away. A light-year is about 9.5 trillion kilometers.

    Researchers have named the planet Kepler-90i. Like Earth, Kepler-90i is the third farthest planet from its sun.

    However, Kepler-90i is much closer to its sun. It only takes the planet 14 days to orbit Kepler-90. So, its surface is much warmer -- 427 degrees Celsius. In fact, all the planets in the Kepler-90 solar system orbit closer to their sun than Earth does to our sun.

    So far, this is the only other eight-planet solar system that researchers have found. Eight is the largest number of planets ever observed around a single sun.

    Our solar system had nine planets up until 2006, when the International Astronomical Union decided Pluto did not meet the requirements to be considered a planet. Instead, the group renamed it a dwarf planet.

    But some astronomers believe there could be a large ninth planet far off in our solar system. They call it Planet X, and believe it is the size of Neptune.

    Researchers also believe there could be nine or more planets in the Kepler-90 solar system.

    Google used data from NASA's special planet-hunting device, called the Kepler Space Telescope, to locate Kepler-90i. The company used the data to develop a computer program with machine learning. This means it can learn and improve itself without a programmer telling it to do so. The program carefully studies planetary signals that are so weak it would take humans years to examine them.

    Christopher Shallue is a senior software engineer at Google in Mountain View, California. He said, "This is a really exciting discovery, and we consider it to be a successful proof of concept to be using neural networks to identify planets, even in ... situations where the signals are very weak."
    克里斯托弗·谢纽(Christopher Shallue)是加州山景城谷歌公司的一名高级软件工程师。他说:“这是一项非常令人兴奋的发现,我们认为它成功证明了利用神经网络识别行星的这一概念,哪怕行星信号非常微弱。”

    NASA astrophysicist and Kepler project scientist Jessie Dotson said she is "so excited to see where this goes next."
    美国宇航局和开普勒项目科学家杰西·朵森(Jessie Dotson)表示她很高兴看到它下一步的发展方向。

    Shallue partnered with astronomer Andrew Vanderburg of the University of Texas at Austin to develop this machine-learning program. The two trained a computer to identify planets beyond our solar system. To do so, it used observations of the minor changes in the brightness of stars when planets passed in front of them that the Kepler Space Telescope had recorded.
    谢纽和德克萨斯大学奥斯丁分校的天文学家安德鲁·范德伯格(Andrew Vanderburg)合作开发了这个机器学习程序。他们两人训练了一台电脑来识别太阳系以外的行星。为了实现这点,它利用了开普勒太空望远镜记录下的恒星在行星经过时亮度细微变化的观测结果。

    Shallue and Vanderburg plan to continue hunting for new planets. They plan to use the program to examine the more than 150,000 stars that the Kepler Space Telescope has already identified.

    So far, Researchers have confirmed the existence of more than 3,560 planets beyond our solar system. The Kepler Space Telescope, which launched in 2009, located about two-thirds of them.

    Another 4,500 possible exoplanets await confirmation.

    I'm Pete Musto.