Indian astronomers have invented a novel strategy for detecting potentially habitable planets with a high likelihood using an Artificial Intelligence-based system.
Humans have been peering towards the universe since the dawn of time, convinced that there are other inhabited worlds out there. According to current estimations, the galaxy's total number of planets is in the billions, possibly exceeding the number of stars. The obvious question is if there are additional planets that could support life, and if there is a technique to anticipate which exoplanets could support life.
Astronomers from the Indian Institute of Astrophysics, an autonomous institute of the Department of Science and Technology, and astronomers from BITS Pilani's Goa campus have developed a new strategy — an anomaly detection method — to find possibly habitable planets with a high likelihood. The technique is predicated on the assumption that the earth is an outlier, with the possibility of a few other outliers among thousands of data points. The research was published in the Royal Astronomical Society's Monthly Notices magazine (MNRAS).
According to the study, out of around 5000 confirmed planets, there are 60 potentially habitable planets and nearly 8000 candidate planets. The assessment is based on their striking resemblance to the planet. These planets could be considered exceptional cases among a large number of 'non-habitable' exoplanets.
"It is defined as an abnormality that Earth is the sole habitable planet among thousands of worlds." Doctor Snehanshu Saha of BITS Pilani K K Birla Goa Campus and Dr. Margarita Safonova of Indian Institute of Astrophysics investigated whether similar 'anomaly candidates' could be discovered using novel anomaly detection technologies.
The fulcrum of the hypothesis that (possibly) habitable exoplanets are anomalies, according to the IIA team, revolves around the well-known anomaly identification challenge in predictive maintenance of industrial systems.
Because the anomaly detector in both cases is working with "imbalanced" data, where the anomalies (number of habitable exoplanets or abnormal behavior of industrial components) are outliers, an anomaly detection technique suitable for industrial systems also works for habitable planet discovery. In comparison to the typical data, there are much less of these.
With so many exoplanets identified, identifying those rare aberrant occurrences by characterizing them in terms of planetary properties, types, populations, and, ultimately, habitability potential, necessitates the knowledge of many planetary parameters gleaned from observations.
This necessitates hours of costly telescope time. Manually scanning thousands of planets and identifying planets that are possibly comparable to Earth is a time-consuming task. Artificial Intelligence (AI) can be useful in the search for habitable worlds.
Researchers developed a novel Artificial Intelligence-based algorithm to detect anomalies and extended it to an unsupervised clustering algorithm to use it to identify the likely habitable exoplanets from exoplanet datasets under the supervision of Prof. Snehanshu Saha of BITS Pilani Goa Campus and Dr Margarita Safonova of Indian Institute of Astrophysics (IIA), Bengaluru.
Proffessor Santonu Sarkar, Jyotirmoy Sarkar, a doctorate student, and Kartik Bhatia, an undergraduate student, all from BITS Pilani Goa Campus, were also part of the research team.
The Multi-Stage Memetic Binary Tree Anomaly Identifier (MSMBTAI) is an AI-based approach that is built on a revolutionary multi-stage memetic algorithm (MSMA). MSMA makes use of the term "meme," which refers to an idea or piece of knowledge that is passed down through the generations through imitation.
A meme denotes future cross-cultural evolution and, as a result, can trigger new learning mechanisms as generations pass. The method can be used as a rapid screening tool for assessing the habitability of properties.
The study used the proposed technique to identify a few planets with abnormal traits comparable to Earth, and the results are reasonably excellent, confirming what astronomers assume. Surprisingly, although this method did not employ surface temperature as a characteristic, it produced similar results in terms of abnormal candidate detection as when it did. This will make future exoplanet research much easier.
You must be logged in to post a comment.