Rubin-LSST

Abstract:

In the past two decades, high-amplitude electromagnetic outbursts have been detected from dormant galaxies and often attributed to the tidal disruption of a star by the central black hole1,2. X-ray emission from the Seyfert 2 galaxy GSN 069 (2MASX J01190869-3411305) at a redshift of z = 0.018 was first detected in July 2010 and implies an X-ray brightening by a factor of more than 240 over ROSAT observations performed 16 years earlier3,4. The emission has smoothly decayed over time since 2010, possibly indicating a long-lived tidal disruption event5. The X-ray spectrum is ultra-soft and can be described by accretion disk emission with luminosity proportional to the fourth power of the disk temperature during long-term evolution. Here we report observations of quasi-periodic X-ray eruptions from the nucleus of GSN 069 over the course of 54 days, from December 2018 onwards. During these eruptions, the X-ray count rate increases by up to two orders of magnitude with an event duration of just over an hour and a recurrence time of about nine hours. These eruptions are associated with fast spectral transitions between a cold and a warm phase in the accretion flow around a low-mass black hole (of approximately 4 × 105 solar masses) with peak X-ray luminosity of about 5 × 1042 erg per second. The warm phase has kT (where T is the temperature and k is the Boltzmann constant) of about 120 electronvolts, reminiscent of the typical soft-X-ray excess, an almost universal thermal-like feature in the X-ray spectra of luminous active nuclei6,7,8. If the observed properties are not unique to GSN 069, and assuming standard scaling of timescales with black hole mass and accretion properties, typical active galactic nuclei with higher-mass black holes can be expected to exhibit high-amplitude optical to X-ray variability on timescales as short as months or years9.

Abstract:

This paper demonstrates a novel and efficient unsupervised clustering method with the combination of a self-organizing map (SOM) and a convolutional autoencoder. The rapidly increasing volume of radio-astronomical data has increased demand for machine-learning methods as solutions to classification and outlier detection. Major astronomical discoveries are unplanned and found in the unexpected, making unsupervised machine learning highly desirable by operating without assumptions and labeled training data. Our approach shows SOM training time is drastically reduced and high-level features can be clustered by training on auto-encoded feature vectors instead of raw images. Our results demonstrate this method is capable of accurately separating outliers on a SOM with neighborhood similarity and K-means clustering of radio-astronomical features. We present this method as a powerful new approach to data exploration by providing a detailed understanding of the morphology and relationships of Radio Galaxy Zoo (RGZ) data set image features which can be applied to new radio survey data.