The SETI Institute has become the first to apply artificial intelligence to the real-time direct detection of faint radio signals from space, marking a significant advancement in the search for extraterrestrial intelligence and the study of rare astronomical phenomena.
The institute, which operates the Allen Telescope Array in Northern California, collaborated with NVIDIA to develop this cutting-edge technology. Using NVIDIA's Holoscan, a sensor processing platform, and the IGX edge computing platform, researchers were able to process real-time data from scientific instruments at unprecedented speeds.
In a pilot project, the team linked their system to 28 antennas pointed at the Crab Nebula. Over 15 hours, they gathered and analysed in real-time more than 90 billion data packets across a 5GHz spectrum, achieving speeds of nearly 100Gbps - twice their previous capability.
Luigi Cruz, a staff engineer at the SETI Institute, developed the real-time data reception and inference pipeline, while Peter Ma, a Breakthrough Listen collaborator, built and trained an AI model to detect fast radio bursts. Wael Farah, Allen Telescope Array project scientist, provided key scientific contributions.
The SETI Institute plans to scale up this pilot software and deploy it across all their radio telescopes at a dozen sites worldwide. They also aim to share this capability with astronomers globally, potentially revolutionising how radio astronomy is conducted.