A new ML algorithm has confirmed the presence of 50 new planets. When fed with data about potential planets, the algorithm can calculate the likelihood that it's a false positive or a true planet. The planets it confirmed range in size from Earth to Neptune, with orbits between one and 200 days.
- The system is based on data collected by NASA's Kepler space telescope, which can look for signs of planets passing in front of stars. It comes from Warwick's Departments of Physics and Computer Science and the Alan Turing Institute.
- Warwick's David Armstrong, whose focus is applying machine learning to astrophysical problems, said no one had used an ML technique yet to validate planets, though it has been used to rank potential candidates.
- Through the algorithm, researchers can determine what the "precise statistical likelihood is" that it's a true planet. If there's less than a 1% chance it's not, it's considered validated.
- The researchers expect the system could validate thousands of planetary candidates in the future.
- Currently, there are 4,201 confirmed exoplanets.
- You can read the study here.