Astrophysicists have further decoded the secrets of exoplanets using neural networks.
A team of researchers at Ludwig Maximilians University (LMU) in Munich, Germany has made a breakthrough in research into the atmospheres of exoplanets.
Exoplanets are planets outside our solar system that orbit other stars in the Milky Way galaxy or beyond it. Some of them lead rogue lives, traveling alone across space.
Research on distant worlds has been ongoing since the 1990s, but it intensified with the launch of the James Webb Telescope, the largest telescope in space.
A recent discovery by Munich scientists has advanced the study of distant worlds by using physics-informed neural networks (PINNs), allowing for more precise modeling of complex light scattering in exoplanet atmospheres than ever before.
PINNs are a type of artificial intelligence that uses input data and known laws of physics to solve problems and draw conclusions. By combining the data and the knowledge of physics, PINNs can make more accurate predictions, even when there is a lack of input data.
Exoplanet atmospheres hold the secrets
To understand exoplanets, it’s crucial to investigate the light scattering in the atmosphere, particularly the scattering of clouds. When distant exoplanets move in front of their star, they block some of the starlight, and a tiny bit of this light passes through the planet's atmosphere. This causes changes in the light spectrum, showing details about the atmosphere, like its chemical makeup, temperature, and cloud cover.
To study these changes, scientists need powerful models that can quickly create and analyze millions of different light spectra.
According to LMU’s scientists, PINNs create new possibilities for analyzing exoplanet atmospheres that previous models couldn’t, particularly in terms of cloud dynamics, and have the potential to greatly enhance our understanding of these distant worlds.In July, NASA studied the atmosphere of WASP-39 b exoplanet and confirmed the tidally locked exoplanets have eternal mornings and evenings, with half of the unknown worlds shrouded in everlasting darkness and winds blowing at 1000 mph.
Earlier this year, Webb’s telescope helped to examine another exoplanet, LHS 1140 b. The data collected showed that the planet could be the closest habitable world to Earth. According to predictions, the planet is either completely covered in ice or has some parts of a liquid ocean and a stable, cloudy atmosphere.
AI advances exoplanet studies
With the intensifying research on exoplanets, AI is becoming an important tool for studying space. In May, NASA head administrator Bill Nelson pointed out that AI can accelerate the pace of discovery.
Specifically designed AI models allow scientists to simulate star systems and analyze the parameters of how a protoplanetary disc containing planets looks and behaves, for example. In an interview with Cybernews, University of Georgia doctoral student Jason P. Terry explained how he employed an AI model to discover new exoplanets.
Your email address will not be published. Required fields are markedmarked