Machine-learning screening of Parkes and Green Bank radio data for technosignatures in Breakthrough Listen finds no confirmed detections

To the point

Machine learning was used to search about 1e11 spectrograms from Parkes and Green Bank radio data for extraterrestrial technosignatures, ranking candidates by novelty and persistence and vetting around 20,000 checks, but none passed basic scrutiny, highlighting the ongoing difficulty of distinguishing real signals from interference in this open, data‑heavy search led by Snir Pardo, Dovi Poznanski, Steve Croft, Andrew P. V. Siemion, and Matthew Lebofsky.