Searching for extra-solar planets requires more data and time, but the process can be improved through Google's artificial intelligence (AI) system.
Engadget, the Kepler spacecraft of the US Airways and Space Agency spent four years concentrating on a streak of light in the sky, and during that time it collected 14 billion data points from 200,000 spectators. star.By taking this light trail every 30 minutes, Kepler discovered a large star named Kepler 90. Data from NASA revealed that a total of seven planes orbiting Kepler 90 were named. Turns Kepler 90b, 90c, 90d, 90e, 90f, 90g and 90h.But in the latest detection from Google's AI system, led by researcher Chris Shallue in collaboration with Austin University astronomer Andrew Vanderburg, helped discover a new planet flying around Kepler 90 and Named as Kepler 90i.
The AI system used by Shallue and Vanderburg is a mechanistic system that identifies extrasolar planets based on the light data that NASA records. Specifically, they used a set of 15,000 Kepler signals, known from planets or other objects such as stars, to train machine learning systems. When they tested the system, the researchers found that it could accurately discriminate signals from planets for up to 96% improvement.
They then took Kepler's data from 670 known stars, finding that two planets had been ignored by scientists. In addition to Kepler 90i there is Kepler 80g. Note that, with the new findings, the Kepler 90 star becomes a system with eight planets orbiting around very similar to the solar system we live in - which also has eight planets orbiting around.
Shallue and Vanderburg say it is only the beginning, because the group only searches for 670 stars out of 200,000, so many other planets have not yet been collected in Kepler data. New ideas and techniques like machine learning can help discover new objects for years to come.