Helen Qu
Bio: I am a fourth year PhD candidate in Physics at the University of Pennsylvania, advised by Dr. Masao Sako. I work with type Ia supernovae as cosmological probes as part of the Dark Energy Survey Collaboration as well as the Nancy Grace Roman Space Telescope supernova science investigation team. I am particularly interested in applications of machine learning and other modern data science techniques to large-scale astronomical survey data.
Talk Title: A Convolutional Neural Network Approach to Supernova Classification
Talk Abstract: One of the brightest objects in the universe, supernovae are powerful explosions marking the end of a star’s lifetime. Supernova type is defined by spectroscopic emission lines, but obtaining spectroscopy is often logistically unfeasible. Thus, the ability to identify supernovae by type using time-series image data alone is crucial, especially in light of the increasing breadth and depth of upcoming telescopes. We present a convolutional neural network method for fast supernova time-series classification, with observed brightness data smoothed in both the wavelength and time directions with Gaussian process regression. We apply this method to full duration and truncated supernova time-series, to simulate retrospective as well as real-time classification performance. Retrospective classification is used to differentiate cosmologically useful Type Ia supernovae from other supernova types, and this method achieves > 99% accuracy on this task. We are also able to differentiate between 6 supernova types with 60% accuracy given only two nights of data and 98% accuracy retrospectively.