This paper presents the relationship between planetary radii and orbital periods for confirmed transiting and verified planets. CP is indicated by discovery source, Kepler is indicated by ‘+’, K2 by diamond ‘⋄’ and TESS by ‘*’. Verified planets boosted with ExoMiner V1.2 are indicated by magenta squares. For reference, the earth is shown as a black circle. — astro-ph.EP
In a breakthrough, a team of machine learning scientists and astronomers from the University Space Research Association (USRA), SETI Institute, and NASA discover 69 new exoplanets using advanced machine learning techniques Did.
This discovery has been accepted for publication in the Journal of Astronomy. This important advance has been made possible by harnessing the power of artificial intelligence, which promises to broaden our understanding of the universe and pave the way for future discoveries.
Exoplanets and planets outside our solar system have long been of interest to scientists as well as the general public. A variety of approaches have been used to discover exoplanets, including the transit method, which has led to the discovery of most exoplanets. For example, Kepler and his TESS mission effort were based on a transit method that monitors whether a star of interest periodically dims in brightness, known as a transit event. However, not all detections by transiting events are exoplanets, and may be due to various sources of false positives, such as eclipsing binaries.
Traditionally, complementary observations were used to rule out false positives to ensure that signals detected as exoplanets were not due to false positive sources. But in contrast, statistical and machine learning techniques are advancing and relying on a new process called ‘validation’ that has been developed to discover new exoplanets. Rather than relying on new observations to complement the transit method, the 69 newly discovered exoplanets are validated using a previously developed deep learning called ExoMiner and a concept called multiplicity. Astronomers strongly believe that the reliability of newly detected signals around stars that already have exoplanets is much higher than for stars without exoplanets, because multiplicity increases the probability. I’m here.
Dr. Hamed Valizadegan, a machine learning scientist at the University Space Research Association and lead author of the paper, said: Eliminate false positives and validate 69 new exoplanets. His 69 newly discovered exoplanets vary widely in size, orbital period, proximity to their star, and other characteristics, and may improve our understanding of the number of exoplanets in the universe. . “
The team recently developed a new deep neural network, ExoMiner, which will be used to validate 301 new exoplanets in 2021. However, existing transit signal classifiers, including ExoMiner, do not use information about the composition of the planetary system, such as the number of known planets or false positive signals. The configuration of the system can be used to improve the reliability of the classifier for validating new exoplanets.
Existing verification methods ignore multiplicity boost information. In the latest research, the team used the multiplicity boost framework proposed for ExoMiner V1.2. It addressed some of the shortcomings of the original ExoMiner classifier (Valizadegan et al. 2022) and validated 69 new exoplanets against systems with multiple KOIs. Kepler Catalog.
The discovery of 69 new exoplanets using machine learning is a pivotal milestone in exploratory research, bringing us closer to answering fundamental questions about our place in the universe. As we continue to explore the vast depths of space, the collaboration of astronomy and artificial intelligence is expected to redefine our understanding of the universe.
Additional resources:
“Multiplicity Boosting of Passing Signal Classifiers: Validation of 69 New Exoplanets Using ExoMiner’s Multiplicity Boosting”
team members, collaborators, customers
Hamed Varizadegan (USRA), Miguel Martinho (USRA), John M. Jenkins (NASA Ames), Douglas A. Caldwell (SETI Research Institute), Joseph D. Twicken (SETI Research Institute), Stephen T. T. Bryson (NASA Ames)
Hamed Valizadegan and Miguel Martinho are supported through NASA NAMS Contract NNA16BD14C, TESS GI Cycle 4 Contract 80NSSC22K0184, and NASA ROSES XRP Proposal 22-XRP22_2-0173. Douglas Caldwell and Joseph Twicken are supported through NASA Cooperation Agreement 80NSSC21M0079.
About USRA
Founded in 1969 at the request of the United States Government and under the auspices of the National Academy of Sciences, the University Space Research Association (USRA) is a nonprofit corporation dedicated to advancing space-related science, technology, and engineering. USRA operates scientific institutions and facilities and conducts other major research and educational programs. USRA works with the university community and employs in-house scientific leadership, innovative research and development, and project management expertise. For more information on USRA, please visit www.usra.edu.
space biology
