FDL 2018 - Exoplanets - Classification of Planet Candidates in TESS Mission Data with Deep Neural Nets

FRONTIER DEVELOPMENT LAB 2018: EXOPLANETS Classification of Planet Candidates in TESS Mission Data with Deep Neural Nets Megan Ansdell, Yani Ioannou, Hugh Osborn, Michele Sasdelli PROBLEM SOLUTION Astronet (Shallue & Vanderburg 2017) on Kepler light curves... More

FRONTIER DEVELOPMENT LAB 2018: EXOPLANETS Classification of Planet Candidates in TESS Mission Data with Deep Neural Nets Megan Ansdell, Yani Ioannou, Hugh Osborn, Michele Sasdelli PROBLEM SOLUTION Astronet (Shallue & Vanderburg 2017) on Kepler light curves Classifying Candidate Transits in Processed Light Curves from Machine Learning to Increase the Efficacy and Example Exoplanet Transits Kepler and TESS with Deep Neural Nets Yield of Exoplanet Transit Detections This approach is based on Shallue & Vanderburg (2017), who used a deep convolutional neural network called Astronet to classify planet candidate signals identified by the Kepler pipeline. Exoplanets can be detected using the transit technique, which works by The network was a binary classifier, which identified a signal as a true transiting exoplanet or a measuring the brightness of a target star as a function of time, producing false positive caused by astrophysical or instrumental phenomena. We improved this method a flux ti Less