An artificial neural network to discover hypervelocity stars: candidates in Gaia DR1/TGAS

Abstract

The paucity of hypervelocity stars (HVSs) known to date has severely hampered their potential to investigate the stellar population of the Galactic Centre and the Galactic Potential. The first Gaia data release (DR1, 2016 September 14) gives an opportunity to increase the current sample. The challenge is the disparity between the expected number of hypervelocity stars and that of bound background stars. We have applied a novel data mining algorithm based on machine learning techniques, an artificial neural network, to the Tycho-Gaia astrometric solution (TGAS) catalogue. With no pre-selection of data, we could exclude immediately ∼ 99% of the stars in the catalogue and find 80 candidates with more than 90% predicted probability to be HVSs, based only on their position, proper motions, and parallax. We have cross-checked our findings with other spectroscopic surveys, determining radial velocities for 30 and spectroscopic distances for 5 candidates. In addition, follow-up observations have been carried out at the Isaac Newton Telescope for 22 stars, for which we obtained radial velocities and distance estimates. We discover 14 stars with a total velocity in the Galactic rest frame > 400 km s −1 , and 5 of these have a probability > 50% of being unbound from the Milky Way. Tracing back their orbits in different Galactic potential models we find one possible unbound HVS with v ∼ 520 km s −1 , 5 bound HVSs, and, notably, 5 runaway stars with median velocity between 400 and 780 km s −1 . At the moment, uncertainties in the distance estimates and ages are too large to confirm the nature of our candidates by narrowing down their ejection location, and we wait for future Gaia releases to validate the quality of our sample. This test successfully demonstrates the feasibility of our new data mining routine.

Predicting the hypervelocity star population in Gaia

“…We find a total of 0.46 HVSs surviving this magnitude cut. This result is consistent with results in Marchetti et al (2017), which find only one star with both a predicted probability > 50% of being unbound from the Galaxy and a trajectory consistent with coming from the GC.…”

Rates of Stellar Tidal Disruption

“…This moment in time is especially propitious because of the ongoing data releases by the astrometric Galactic survey Gaia that have intensified searches for HVSs (e.g. Marchetti et al 2017Marchetti et al , 2018Boubert et al 2018;Bromley et al 2018). Likewise, a comparison between rates of single and double TDE in other galaxies could provide an extra consistency check on the relationship between single and binary star loss cones.…”

Surveying the reach and maturity of machine learning and artificial intelligence in astronomy

“…With regards to the role of ML and AI in advancing knowledge in astronomy, there was clear evidence from the sample of recent publications that discovery tasks are being performed with all of the data types: images (Ciuca & Hernández, ; Gomez Gonzalez, Absil, & Van Droogenbroeck, ; Hartley, Flamary, Jackson, Tagore, & Metcalf, ; Jacobs et al, ; Lanusse et al, ; Morello, Morris, Van Dyk, Marston, & Mauerhan, ; Pourrahmani et al, ; Wan et al, ); spectroscopy (Bu, Lei, Zhao, Bu, & Pan, ; Li et al, ); photometry (Ostrovski et al, ; Timlin et al, ; Vida & Roettenbacher, ); light curves (Armstrong et al, ; Cohen et al, ; Giles & Walkowicz, ; Hedges, Hodgkin, & Kennedy, ; Heinze et al, ; Peña et al, ; van Roestel et al, ); time‐series (Connor & van Leeuwen, ; Farah et al, ; Michilli et al, ; Morello et al, ; Pang et al, ; Tan et al, ); catalogues (Lin et al, ; Marchetti et al, ; Nguyen, Pankratius, Eckman, & Seager, ; Yan et al, ); and simulation (Kuntzer & Courbin, ; Nadler et al, ; Xu & Offner, ).…”

“…Two key activities in stellar astronomy are spectral classification (e.g., Garcia‐Dias, Allende Prieto, Sánchez Almeida, & Ordovás‐Pascual, ; Wang, Guo, & Luo, with k ‐means clustering; Kong et al, ; classification of young stellar objects with eight different methods by Miettinen, ) and photometric classification (e.g., Ksoll et al, ; Zhang et al, with SVM, RF, and Fast Boxes). Many new examples of specific stellar classes have been discovered, such as Wolf‐Rayet stars (Morello et al, ), blue horizontal branch stars (Wan et al, ), hot sub dwarf stars (Bu et al, ), and rare hypervelocity stars (Marchetti et al, ). ML/AI have also led to the discovery of unresolved binary stars in simulated catalogues using RF and ANN algorithms (Kuntzer & Courbin, ), and new pulsars, and fewer false positives, from the LOFAR Tied‐Array All‐Sky Survey (Michilli et al, ; Tan et al, ).…”