Many people, including great scientists like Stephen Hawking, are afraid of how Artificial Intelligence will evolve in the near future. Whether this technology will exterminate us one day or not, the truth is that it continues for the moment to be extremely useful. One recent proof is what a team of scientists from Warwick University (Coventry, England) managed to do with the help of a machine learning algorithm.
After receiving huge data from NASA’s Kepler mission and the Transiting Exoplanet Survey Satellite (TESS), the machine learning algorithm was able to confirm the existence of 50 new exoplanets. These cosmic objects present themselves in various sizes – from smaller than Earth to sizes larger than Neptune.
Training the algorithm is still required
The research team had to train the algorithm by teaching it how to differentiate between confirmed planets and false positives.
David Armstrong, who is the lead author of the study and working at the University of Warwick department of physics, declared:
Rather than saying which candidates are more likely to be planets, we can now say what the precise statistical likelihood is,
Where there is less than a 1 percent chance of a candidate being a false positive, it is considered a validated planet.
However, Armstrong still admits the following:
We still have to spend time training the algorithm, but once that is done it becomes much easier to apply it to future candidates,
Whether astronomers like to admit it or not, the techniques for finding exoplanets are pretty scarce. Exoplanets are mostly found when they pass in front of their host stars. Otherwise, those planets are usually far too weakly illuminated in order for the scientists to detect them. Therefore, the researchers from Warwick University hope that the AI will open new paths towards planet validation techniques.
The story was published in the journal Monthly Notices of the Royal Astronomical Society.
