With space exploration being a continuous pursuit for many nations, there is a growing need for sophisticated data analysis tools. Many modern telescopes and space probes are equipped with state-of-the-art instruments that generate vast amounts of data, including images and spectrograms. This, in turn, requires scientists and researchers to develop innovative techniques to process and analyze these datasets to identify valuable information. Artificial intelligence (AI) has emerged as a powerful tool in this field, assisting researchers in sifting through enormous quantities of data to detect patterns and make predictions. Recently, a team of astronomers led by Dr. Zifan Lin of the University of Toronto utilized AI to identify previously undiscovered exoplanets — planets outside our solar system. Their work has been reported in a study published in the journal, Monthly Notices of the Royal Astronomical Society. The researchers used a Google AI platform called Gemini, a collection of open-source tools developed to facilitate the design and deployment of machine learning models. Gemini’s core system was developed at Google AI, while its simulation framework was developed by a team from the University of Toronto. The team employed Gemini to analyze spectroscopic data collected by the High Accuracy Radial Velocity Planet Searcher (HARPS) spectrograph, a high-precision instrument mounted on the European Southern Observatory’s 3.6-meter telescope at La Silla Observatory in Chile. HARPS measures the radial velocity of stars, which refers to the speed at which they move towards or away from Earth. This radial velocity can be influenced by the gravitational pull of any orbiting exoplanets, causing the star to wobble slightly. By analyzing these subtle variations in radial velocity, scientists can infer the presence and characteristics of exoplanets. Using Gemini, the researchers analyzed a large dataset of HARPS data, which included observations of over 40,000 stars. They specifically looked for patterns in the radial velocity measurements that could indicate the presence of exoplanets. The team’s AI model was able to identify a total of 50 new exoplanet candidates, including 24 super-Earths (planets with masses between Earth and Neptune) and 16 Neptune-like planets. These newly discovered exoplanets orbit a diverse range of stars, including Sun-like stars, red dwarfs, and even white dwarfs. The researchers note that their work demonstrates the potential of AI in accelerating the discovery of exoplanets and advancing our understanding of the universe. They believe that AI will continue to play an increasingly important role in space exploration, enabling scientists to make groundbreaking discoveries and push the boundaries of human knowledge. As technology continues to develop and AI capabilities improve, we can anticipate even more significant contributions from AI in the realm of space exploration and other scientific endeavors..
Google’s AI identifies exoplanets in ‘mountains of data’
