With the help of an Artificial Intelligence based algorithm, a team of Indian Astronomers belonging to the Indian Institute of Astrophysics, along with the help of Astronomers from BITS Pilani, Goa Campus has devised a new method, an anomaly detection method, which can help to identify potentially habitable planets with high probability. The Indian Institute of Astrophysics is an autonomous institute of the Department of Science and Technology.
The official release from the Ministry of Science and Technology stated that the base of the approach is the hypothesis that the Earth is an anomaly with the possibility of the existence of a few other anomalies among thousands of data points. According to the study, out of the 5000 confirmed and around 8,000 candidate planets proposed, there are 60 potentially habitable planets. The assessment is based on the planets close similarity to earth. These planets can be viewed as potential candidates for anomalous instances in a huge pool of `non-habitable’ exoplanets.
Till now, Earth is the only habitable planet among thousands of planets and so is defined as an anomaly or peculiar. The astronomers explored to see whether such peculiar candidates can be found using novel anomaly detection methods. “Since it is the only habitable planet among thousands of planets, the Earth is defined as an anomaly. We explored whether similar differentiating or anomalous candidates can be found using novel anomaly detection methods,” Dr Margarita Safonova of Indian Institute of Astrophysics and Dr Snehanshu Saha of BITS Pilani said.
The Indian Institute of Astrophysics team explained that the fulcrum of the idea that assumes potentially habitable exoplanets as anomalies revolves around the well-known anomaly detection method in predictive maintenance of industrial systems. The anomaly detection technique used for the prediction and maintenance of industrial systems applies equally well for the detection of a habitable planet since in both cases, the inconsistency detector is dealing with imbalanced or inconsistent data, where the anomalies (anomalous behaviour of industrial components or in the present case the number of habitable exoplanets) are deviations or irregularity. These irregularities or anomalies are far less in number when compared to the normal data.
However, finding those rare anomalous instances in an extremely large number of discovered exoplanets, that too by characterizing them in terms of planetary parameters, types, populations, and ultimately by their habitability potential, requires the knowledge of multiple planetary parameters from observations, which, in turn, requires hours of costly telescope time. Scanning thousands of planets manually and identifying planets that are potentially similar to Earth is an exhaustive task to do. Here comes the application of Artificial Intelligence as it can be utilized effectively to find potentially habitable planets.
Under the supervision of Prof. Snehanshu Saha of BITS Pilani, Goa and Dr Margarita Safonova of the Indian Institute of Astrophysics, Bengaluru the researchers developed an innovative Artificial Intelligence based algorithm to detect anomalies. They extended it to an unsupervised clustering algorithm so as to use it to identify the potentially habitable exoplanets from the exoplanet datasets.
How the Artificial Intelligence functions
Multi-Stage Memetic Binary Tree Anomaly Identifier (MSMBTAI), the Artificial Intelligence-based method is based on a unique multi-stage memetic algorithm (MSMA). Mimetic Algorithm (MSMA) uses the generic notion of a meme, which is, an idea or knowledge that gets transferred from one person to another by imitation or mimic. Memetic algorithms are a class of metaheuristic algorithms, where an evolutionary approach is hybridised with problem-specific information. The objective of the hybridisation is to accelerate the discovery of good solutions, for which the Evolutionary Algorithm (EA) would have taken a long time to reach, or it was never possible to reach.
A meme indicates the cross-cultural evolution of descendants and, so can produce new learning mechanisms as the generations pass. The MSMA algorithm can act as a quick screening tool to evaluate the habitability perspectives of the planets from the observed properties.
Through the proposed technique, the study identified some planets which demonstrate similar anomalous characteristics as that of Earth. In agreement with what astronomers believe, the algorithm delivered reasonably good results. Interestingly, the method delivered similar results in terms of anomalous candidate detection when it did not use surface temperature as a feature, compared to when it did use surface temperature as a feature.
