Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Pan-STARRS1 has carried out a set of distinct synoptic imaging sky surveys including the $3\pi$ Steradian Survey and the Medium Deep Survey in 5 bands ($grizy_{P1}$). The mean 5$\sigma$ point source limiting sensitivities in the stacked 3$\pi$ Steradian Survey in $grizy_{P1}$ are (23.3, 23.2, 23.1, 22.3, 21.4) respectively. The upper bound on the systematic uncertainty in the photometric calibration across the sky is 7-12 millimag depending on the bandpass. The systematic uncertainty of the astrometric calibration using the Gaia frame comes from a comparison of the results with Gaia: the standard deviation of the mean and median residuals ($ \Delta ra, \Delta dec $) are (2.3, 1.7) milliarcsec, and (3.1, 4.8) milliarcsec respectively. The Pan-STARRS system and the design of the PS1 surveys are described and an overview of the resulting image and catalog data products and their basic characteristics are described together with a summary of important results. The images, reduced data products, and derived data products from the Pan-STARRS1 surveys are available to the community from the Mikulski Archive for Space Telescopes (MAST) at STScI.

The Pan-STARRS1 Surveys

We present observations of SN 2015bn (=PS15ae = CSS141223-113342+004332 = MLS150211-113342+004333), a Type I superluminous supernova (SLSN) at redshift z = 0.1136. As well as being one of the closest SLSNe I yet discovered, it is intrinsically brighter (M_U ≈ -23.1) and in a fainter galaxy (M_B ≈ -16.0) than other SLSNe at z ~ 0.1. We used this opportunity to collect the most extensive data set for any SLSN I to date, including densely sampled spectroscopy and photometry, from the UV to the NIR, spanning −50 to +250 days from optical maximum. SN 2015bn fades slowly, but exhibits surprising undulations in the light curve on a timescale of 30–50 days, especially in the UV. The spectrum shows extraordinarily slow evolution except for a rapid transformation between +7 and +20–30 days. No narrow emission lines from slow-moving material are observed at any phase. We derive physical properties including the bolometric luminosity, and find slow velocity evolution and non-monotonic temperature and radial evolution. A deep radio limit rules out a healthy off-axis gamma-ray burst, and places constraints on the pre-explosion mass loss. The data can be consistently explained by a ≳ 10 M_☉ stripped progenitor exploding with ~ 10^(51) erg kinetic energy, forming a magnetar with a spin-down timescale of ~20 days (thus avoiding a gamma-ray burst) that reheats the ejecta and drives ionization fronts. The most likely alternative scenario—interaction with ~20 M_☉ of dense, inhomogeneous circumstellar material—can be tested with continuing radio follow-up.

/pdf/sn-2015bn-a-detailed-multi-wavelength-view-of-a-nearby-4498ubzfg0.pdf

SN 2015bn: A DETAILED MULTI-WAVELENGTH VIEW OF A NEARBY SUPERLUMINOUS SUPERNOVA

Superluminous supernovae are a new class of supernovae that were recognized about a decade ago. Both observational and theoretical progress has been significant in the last decade. In this review, we first briefly summarize the observational properties of superluminous supernovae. We then introduce the three major suggested luminosity sources to explain the huge luminosities of superluminous supernovae, i.e., the nuclear decay of 56Ni, the interaction between supernova ejecta and dense circumstellar media, and the spin down of magnetars. We compare these models and discuss their strengths and weaknesses.

Superluminous supernovae

We investigate the light-curve properties of a sample of 26 spectroscopically confirmed hydrogen-poor superluminous supernovae (SLSNe-I) in the Palomar Transient Factory survey. These events are brighter than SNe Ib/c and SNe Ic-BL, on average, by about 4 and 2 mag, respectively. The peak absolute magnitudes of SLSNe-I in rest-frame g band span −22 ≾ M g ≾ −20 mag, and these peaks are not powered by radioactive ^(56)Ni, unless strong asymmetries are at play. The rise timescales are longer for SLSNe than for normal SNe Ib/c, by roughly 10 days, for events with similar decay times. Thus, SLSNe-I can be considered as a separate population based on photometric properties. After peak, SLSNe-I decay with a wide range of slopes, with no obvious gap between rapidly declining and slowly declining events. The latter events show more irregularities (bumps) in the light curves at all times. At late times, the SLSN-I light curves slow down and cluster around the ^(56)Co radioactive decay rate. Powering the late-time light curves with radioactive decay would require between 1 and 10 M⊙ of Ni masses. Alternatively, a simple magnetar model can reasonably fit the majority of SLSNe-I light curves, with four exceptions, and can mimic the radioactive decay of ^(56)Co, up to ~400 days from explosion. The resulting spin values do not correlate with the host-galaxy metallicities. Finally, the analysis of our sample cannot strengthen the case for using SLSNe-I for cosmology.

/pdf/light-curves-of-hydrogen-poor-superluminous-supernovae-from-262cf3xp5d.pdf

Light curves of hydrogen-poor superluminous supernovae from the palomar transient factory

We present two hydrogen-rich superluminous supernovae (SLSNe): SN2103hx and PS15br. These objects, together with SN2008es, are the only SLSNe showing a distinct, broad H α feature during the photospheric phase; also, they show no sign of strong interaction between fast moving ejecta and circumstellar shells in their early spectra. Despite the fact that the peak luminosity of PS15br is fainter than that of the other two objects, the spectrophotometric evolution is similar to SN2103hx and different from any other supernova in a similar luminosity space. We group all of them as SLSNe II and hence they are distinct from the known class of SLSN IIn. Both transients show a strong, multicomponent H α emission after 200 d past maximum, which we interpret as an indication of the interaction of the ejecta with an asymmetric, clumpy circumstellar material. The spectra and photometric evolution of the two objects are similar to Type II supernovae, although they have much higher luminosity and evolve on slower time-scales. This is qualitatively similar to how SLSNe I compare with normal type Ic, in that the former are brighter and evolve more slowly. We apply a magnetar and an interaction semi-analytical code to fit the light curves of our two objects and SN2008es. The overall observational data set would tend to favour the magnetar, or central engine, model as the source of the peak luminosity, although the clear signature of late-time interaction indicates that interaction can play a role in the luminosity evolution of SLSNe II at some phases.

/pdf/on-the-nature-of-hydrogen-rich-superluminous-supernovae-4faspw6uru.pdf

M. Wilman

Papers

On the nature of hydrogen-rich superluminous supernovae