Day 1043

Thursday.


Nature 529 484-489
Mastering the game of Go with deep neural networks and tree search
Silver et al

The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves.  Introduce a new approach to computer Go that uses 'value networks' to evaluate board positions and 'policy networks' to select moves.  These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play.  Without any lookahead search, the neural networks play Go at the level of state-of-the-art Monte Carlo tree search programs that simulate thousands of random games of self-play.  Also introduce a new search algorithm that combines Monte Carlo sim with value and policy networks.  Using this search algorithm, the program AlphaGo achieved a 99.8% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0.  This is the first time that a computer program has defeated a human professional pleaser in the full-sized game of Go, a feat previously thought to be at least a decade away.

1601.07178
The feeble giant.  Discovery of a large and diffuse Milky Way dwarf galaxy in the constellation of Crater
Torrealba et al

Discovery of the Crater 2 dwarf galaxy, identified in imaging data of the VST ATLAS survey.  Half-light radius ~1100 pc, 4th largest dwarf in the MW, surpassed only by the LMC, SMC nd the Sgr dwarf.  With a total luminosity of M_V~-8, this satellite galaxy is also is one of the lowest surface brightness dwarfs.  Falling under the nominal detection boundary of 30 mag arcsec^-2, it compares in nebulosity to the recently discovered Tuc 2 and Tuc IV and UMa II.  Crater 2 is located ~120 kpc from the Sun and appears to be aligned in 3D with the enigmatic globular cluster Crater, the pair of ultra-faint dwarfs Leo IV and Leo V and the classical dwarf Leo II.  Argue that such arrangement is probably not accidental and, in fact, can be viewed as the evidence for the accretion of the Crater-Leo group.


1601.07182
Detection and removal of artifacts in astronomical images
Desai, Mohr, Bertin, Kummel, Wetzstein

Astronomical images from topical photometric surveys are typically contaminated with transient artifacts such as CRs, satellite trails and scattered light.  Developed and tested an algorithm that removes these artifacts using a deep, artifact free, static sky coadd image built up through the median combination of PSF homogenized, overlapping single epoch images.  Transient artifacts are detected and based in each single epoch image through comparison with an artifact free, PSF-matched simulated image that is constructed using the PSF-corrected, model fitting catalog from the artifact free coadd image together with the position variable PSF model of the single epoch image.  This approach works well not only for cleaning single epoch images with worse seeing than the PSF homogenized coadd, but also the traditionally much more challenging problem of cleaning single epoch images with better seeing.  In addition to masking transient artifacts, developed an interpolation approach that uses the local PSF and performs well in removing artifacts whose widths are smaller than the PSF full width at half maximum, including CRs, the peaks of saturated stars and bleed trails.  Tested this algorithm on DES SV data and present performance metrics.  More generally, algorithm can be applied to any survey which images the same part of the sky multiple times.