AI in the Search for Extraterrestrial Life: How Machine Learning Filters Radio Signals and Vets Exoplanets

Machine learning is being applied across SETI, exoplanet and biosignature research to reject radio interference, vet planet candidates and estimate atmospheric properties, but none of these systems has established that life exists beyond Earth. The UCLA SETI group's convolutional neural network assesses signal consistency across pointings, while a 2023 Nature Astronomy study by Peter Ma and collaborators combined a variational autoencoder with a random-forest classifier to identify eight signals of interest from 820 stars observed by Breakthrough Listen's Green Bank Telescope, none of which have been redetected in follow-up observations.