|
Abstract:
|
Astronomical transients are stellar objects that become temporarily brighter on various timescales and have led to some of the most significant discoveries in cosmology and astronomy. Some of these transients are supernovae that are useful for measuring the composition and expansion of the universe, some are the visible counterparts of gravitational-wave sources such as neutron star mergers, while others still, are rare, exotic, or entirely new kinds of exciting stellar explosions. Recently, astronomy has entered a new era of big data, recording unprecedented numbers of multi-wavelength transients. To meet this demand, we have developed a novel machine learning approach, RAPID (Real-time Automated Photometric Identification using Deep learning), that automatically classifies transients as a function of time. Using a deep recurrent neural network (RNN) with Gated Recurrent Units (GRUs), we are able to quickly classify multi-channel, sparse, time series datasets into several astrophysical types. We build the deep architecture with Tensorflow and train on realistic data streams from current and future transient surveys.
|
