A group of scientists has developed an automatic artificial intelligence learning model that predicts possible planetary systems With planets similar to Earth, to accelerate the future search for twins of our world.
The Search for exoplanets similar to the Earthplanets that orbit stars other than our sun, is a central theme in current planetary research, since it is very likely that there is extraterrestrial life there.
The first exoplanets were discovered in the 1990s, and since then, we have discovered thousands of them. Thanks to the telescopes and the space missions, we have been able to study the composition, the atmosphere and the conditions of these distant worlds.
In this sense, researchers from the University of Bern have developed with artificial intelligence a Innovative automatic learning model which identifies planetary systems that could house planets similar to the Earth. The study has just been published in Astronomy & Astrophysics.
Automatic learning
An automatic learning model is a Statistical tool which is trained with data to recognize certain types of patterns and perform predictions.
Postdoctoral researcher Jeanne Davault, main author of the study, explains in a statement that “our model is based on an algorithm that I developed and was trained To recognize and classify planetary systems that house planets similar to the Earth».
The model is based on previous studies for infer a correlation between the presence or absence of a planet similar to the Earth and the properties of your system.
Berna model
The algorithm trained and tested with data from the so -called Bern model of planetary training and evolution. «Berna’s model Allows you to obtain information on how the planets were formeds, how they have evolved and what types of planets develop under certain conditions on a disc Protoplanetary», Explains the co -author, Dr. Yann Alibert.
Since 2003, Bern model has continually developed at the University of Bern. «Bern’s model is one of the few mOdelos worldwide that offers such a wide wealth of interrelated physical processes and allows you to carry out a study like the current one, ”continues Alibert.
Impressive results
The algorithm of the new automatic learning model trained and tested using data on synthetic planetary systems of the Berna model. «The results are impressive: the algorithm reaches precision values of up to 0.99, which means that 99 % of the systems identified by the automatic learning model contain at least one planet similar to the Earth», Says Davault.
The model was subsequently applied to observed planetary systems. «The model identified 44 systems with high probability to house planets similar to the Earth not detected. A subsequent study confirmed the theoretical possibility that these systems house a planet similar to Earth, ”explains Davault.
