Self-supervised learningSelf-supervised learning (SSL) is a paradigm in machine learning for processing data of lower quality, rather than improving ultimate outcomes. Self-supervised learning more closely imitates the way humans learn to classify objects. The typical SSL method is based on an artificial neural network or other model such as a decision list. The model learns in two steps. First, the task is solved based on an auxiliary or pretext classification task using pseudo-labels which help to initialize the model parameters.
Stellar black holeA stellar black hole (or stellar-mass black hole) is a black hole formed by the gravitational collapse of a star. They have masses ranging from about 5 to several tens of solar masses. The process is observed as a hypernova explosion or as a gamma ray burst. These black holes are also referred to as collapsars. By the no-hair theorem, a black hole can only have three fundamental properties: mass, electric charge, and angular momentum. The angular momentum of a stellar black hole is due to the conservation of angular momentum of the star or objects that produced it.
AstrometryAstrometry is a branch of astronomy that involves precise measurements of the positions and movements of stars and other celestial bodies. It provides the kinematics and physical origin of the Solar System and this galaxy, the Milky Way. The history of astrometry is linked to the history of star catalogues, which gave astronomers reference points for objects in the sky so they could track their movements. This can be dated back to Hipparchus, who around 190 BC used the catalogue of his predecessors Timocharis and Aristillus to discover Earth's precession.
Time-domain astronomyTime-domain astronomy is the study of how astronomical objects change with time. Though the study may be said to begin with Galileo's Letters on Sunspots, the term now refers especially to variable objects beyond the Solar System. Changes over time may be due to movements or changes in the object itself. Common targets included are supernovae, pulsating stars, novas, flare stars, blazars and active galactic nuclei. Visible light time domain studies include OGLE, HAT-South, PanSTARRS, SkyMapper, ASAS, WASP, CRTS, and in a near future the LSST at the Vera C.
Hubble Ultra-Deep FieldThe Hubble Ultra-Deep Field (HUDF) is a deep-field image of a small region of space in the constellation Fornax, containing an estimated 10,000 galaxies. The original data for the image was collected by the Hubble Space Telescope from September 2003 to January 2004. It includes light from galaxies that existed about 13 billion years ago, some 400 to 800 million years after the Big Bang. The HUDF image was taken in a section of the sky with a low density of bright stars in the near-field, allowing much better viewing of dimmer, more distant objects.
Recessional velocityRecessional velocity is the rate at which an extragalactic astronomical object recedes (becomes more distant) from an observer as a result of the expansion of the universe. It can be measured by observing the wavelength shifts of spectral lines emitted by the object, known as the object's cosmological redshift. Hubble's law is the relationship between a galaxy's distance and its recessional velocity, which is approximately linear for galaxies at distances of up to a few hundred megaparsecs.
