Modern AI is being used for everything from pharmaceutical development to self-driving vehicles. As it turns out, the tech used for the latter is being incorporated into space exploration. In turn, artificial intelligence is being used to discover all-new exoplanets.
Reported by SciTech, techniques used to drive autonomous vehicles are being used for planetary discovery. A team of astronomers from the University of Geneva have partnered with AI company Disaitek to fuel new detections.
How artificial intelligence is finding new exoplanets
Currently, new exoplanets are discovered by manipulating eclipses. As a far away planet moves in front of nearby stars, decreased brightness allows them to be photographed with telescopes. However, that all depends on knowing that planets are there.
To combat this, the team of astronomers are using AI technology. By using AI to predict when and where planets will be intersecting their stars, the team was able to take pictures of exoplanets that were once impossible to detect.
In its current form, the neural network constantly cross-references past data to discover new planets. During its first test, the artificial intelligence program discovered two interstellar bodies: Kepler-1705b and Kepler-1705c.
A great step forward
The success of the team’s AI tools is the start of something great. Researcher Adrien Leleu reveals that “the use of AI, in particular of ‘deep learning’ as in this paper, is becoming increasingly widespread in astrophysics”.
Leleu explains that doing this degree of AI prediction of this scale results in massive data. Predicting new exoplanets created terabytes upon terabytes of data. Furthermore, this is only the start; as data sets get bigger, result sizes explode in storage requirements. Nevertheless, this is a great start for the future of planetary discovery.
For more articles like this, take a look at our News page.