SETI Leverages Off-the-Shelf ML Tools from Nvidia and Google to Uncover a Record FRB Dataset

To the point

Berkeley SETI uses off‑the‑shelf ML tools from Nvidia and Google to sift huge telescope data for signs of intelligent life, aided by a graduate student who found the richest dataset yet—21 fast radio bursts in six hours from Green Bank and 72 more a month later—while Andrew Siemion discussed the work with Pandorabots CEO Lauren Kunze at VB Summit 2018; with about a dozen staff, they rely on commodity hardware and ready-made ML to speed discoveries since building instruments from scratch is costly and quickly outdated, and they expect many more FRBs or even extraterrestrial signals hidden in the data.