Post Content The discovery was made using a new AI system called RAVEN, created by University of Warwick researchers. The system analysed observations from over 2.2 million stars collected by Tess during its first four years in space. (Image: Nasa)
Using data from Nasa’s Transiting Exoplanet Survey Satellite (Tess), scientists have found more than 100 planets previously unknown. These planets are outside of our solar system. The discovery was made possible through artificial intelligence, which analysed existing data.
Alongside the confirmed discoveries, the team also found around 2,000 additional candidate planets. About half of these had not been detected before. This could lead to a substantial rise in the number of identified exoplanets, which currently stands at 6,000.
How AI helped identify hidden planets
The discovery was made using a new AI system called RAVEN, created by University of Warwick researchers. The system analysed observations from over 2.2 million stars collected by Tess during its first four years in space.
Tess works by detecting tiny dips in a star’s brightness, which can occur when a planet passes in front of the star, an event known as a transit. However, not every dip is caused by a planet, making verification difficult.
Also Read | The Big Bang’s ‘missing’ half: Why new research suggests black holes swallowed the universe’s antimatter
RAVEN was trained to distinguish between genuine planetary transits and other events that might look similar.
“This represents one of the best-characterised samples of close-in planets and will help us identify the most promising systems for future study,” team leader Marina Lafarga Magro said in a statement.
Addressing a long-standing challenge
A major challenge in astronomy is confirming whether a detected signal genuinely originates from a planet or from other phenomena, such as two stars orbiting each other.
Story continues below this ad
“The challenge lies in identifying if the dimming is indeed caused by a planet in orbit around the star or by something else, like eclipsing binary stars, which is what RAVEN tries to answer,” RAVEN head developer Andreas Hadjigeorghiou said. “Its strength comes from our carefully created dataset of hundreds of thousands of realistically simulated planets and other astrophysical events that can resemble planets.”
The system does more than just detect signals. It can also analyse and validate them in one process, making it more efficient than previous methods.
Also Read | Galactic cosmic rays: Invisible rain that never stops falling
“RAVEN allows us to analyse enormous datasets consistently and objectively,” researcher David Armstrong said. “Because the pipeline is well-tested and carefully validated, this is not just a list of potential planets – it is also reliable enough to serve as a sample to map the prevalence of different types of planets around sun-like stars.”
New insights into unusual planetary systems
The findings also shed light on how common certain types of planets are. The team found that about 10 per cent of stars similar to the Sun host planets that orbit very closely, completing a full orbit in just 16 days.
Story continues below this ad
At the same time, the study confirmed that Neptune-sized planets are extremely rare in these close orbits. This region is often called the “Neptunian desert”.
“For the first time, we can put a precise number on just how empty this ‘desert’ is,” researcher Kaiming Cui said. “These measurements show that Tess can now match, and in some cases surpass, Kepler for studying planetary populations.”
The results demonstrate the impact of artificial intelligence on space exploration. By using AI, scientists can process vast amounts of data and discover new planets that might have otherwise gone undiscovered.
© IE Online Media Services Pvt Ltd
