Breakthrough Listen Machine Learning code and documentation

Here you can find presentations, code, and documentation from the workshop at UC Berkeley on June 2 and 3 2016, in addition to further development of machine learning on Breakthrough Listen data.

• apf_demud: Ran DEMUD novelty detection on APF spectra
• gbt_demud: Ran DEMUD novelty detection on GBT filterbank data that had been converted into spectra by averaging across a 5-minute window
• gbtfil_demud: Ran DEMUD novelty detection on GBT filterbank data (2D)
• gbt_cluster: Re-sampled GBT data and ran clustering results on extracted features (Umaa and Jasper)
• hitrecords_cluster: Looking for clustering in SETI@home and SERENDIP hit records.