Tuesdays 10:30 - 11:30 | Fridays 11:30 - 12:30
Showing votes from 2017-04-14 12:30 to 2017-04-18 11:30 | Next meeting is Tuesday May 19th, 10:30 am.
We demonstrate that, for the baseline design of the CORE satellite mission, the polarized foregrounds can be controlled at the level required to allow the detection of the primordial cosmic microwave background (CMB) $B$-mode polarization with the desired accuracy at both reionization and recombination scales, for tensor-to-scalar ratio values of $r\gtrsim 5\times 10^{-3}$. Under the assumption of perfect control of lensing effects, CORE would measure an unbiased estimate of $r=5\times 10^{-3}$ with an uncertainty of ${\sigma(r=5\times 10^{-3})=0.4\times 10^{-3}}$ after foreground cleaning. In presence of both gravitational lensing effects and astrophysical foregrounds, the significance of the detection is lowered, with CORE achieving yet a $4\sigma$-measurement of $r=5\times 10^{-3}$ after foreground cleaning and $60$% delensing. We deliberately consider detailed sky simulations based on state-of-the-art CMB observations that consist of CMB polarization with $\tau=0.055$ and tensor-to-scalar values ranging from $r=10^{-2}$ to $10^{-3}$, Galactic synchrotron and thermal dust polarization with variable spectral indices over the sky, polarized anomalous microwave emission, polarized infrared and radio sources, and gravitational lensing effects. Using both parametric and blind approaches, we perform the full component separation and likelihood analysis of the simulations, allowing us to quantify both uncertainties and biases on the reconstructed primordial $B$-modes. We report two sources of potential bias for the detection of the primordial B-modes by future CMB experiments: (i) incorrect foreground models, (ii) averaging of foreground spectral indices by pixellization and beam convolution.
We evaluate the performance of four different machine learning algorithms (ANN, Adaboost, GBC, XGBoost), in the separation of pulsars from radio frequency interference (RFI) and other sources of noise, using a dataset consisting of pulsar candidates obtained from the post-processing of a pulsar search pipeline. This dataset was previously used for cross-validation of the {\tt SPINN}-based machine learning engine, which was used for the re-processing of the HTRU-S survey. We report a variety of quality metrics from all four of these algorithms. We apply a model-independent information theoretic approach to determine the features with the most predictive power, and also compare with the feature importance results from the machine learning algorithms, wherever possible. We find that the RMS distance between the folded profile and sub-integrations is the most important feature in Adaboost and XGBoost. In the case of GBC, we find that the logarithm of the ratio of barycentric period and dispersion measure to be the most important feature. The information theoretic approach to feature importance yields a ranking very well matched to that based on GBC. For all the aforementioned machine learning techniques, we report a recall of 100% with false positive rates of 0.15%, 0.077%, 0.1%, 0.08% for ANN, Adaboost, GBC, and XGBoost respectively. Amongst all four of these algorithms, we find that Adaboost has the minimum overlap between the error rates as a function of threshold for detection of pulsars and RFI, and based on this criterion can be considered to be the best.
We study the growth and saturation of the superradiant instability of a complex, massive vector (Proca) field as it extracts energy and angular momentum from a spinning black hole, using numerical solutions of the full Einstein-Proca equations. We concentrate on a rapidly spinning black hole ($a=0.99$) and the dominant $m=1$ azimuthal mode of the Proca field, with real and imaginary components of the field chosen to yield an axisymmetric stress-tensor, hence spacetime. We find that in excess of $9\%$ of the black hole's mass can be transferred into the field. In all cases studied, the superradiant instability smoothly saturates when the black hole's horizon frequency decreases to match the frequency of the Proca cloud that spontaneously forms around the black hole.
The Milky Way dark matter halo is formed from the accretion of smaller subhalos. These sub-units also harbor stars--typically old and metal-poor--that are deposited in the Galactic inner regions by disruption events. In this Letter, we show that the dark matter and metal-poor stars in the Solar neighborhood share similar kinematics due to their common origin. Using the high-resolution Eris simulation, which traces the evolution of both the dark matter and baryons in a realistic Milky Way analog galaxy, we demonstrate that metal-poor stars are indeed effective tracers for the local, virialized dark matter velocity distribution. The dark matter velocities in the Solar neighborhood can therefore be inferred from observations of the smooth inner halo made by the Sloan Digital Sky Survey. This empirical distribution has a lower peak speed and smaller dispersion than what is typically assumed in the Standard Halo Model, affecting the interpretation of direct detection experiments. Specifically, the bounds on the spin-independent scattering cross section are weakened by nearly an order of magnitude for masses below ~10 GeV. Upcoming data from Gaia will allow us to further refine the expected distribution for the smooth dark matter component, and to test for the presence of local substructure.
Astrophysics and cosmology are rich with data. The advent of wide-area digital cameras on large aperture telescopes has led to ever more ambitious surveys of the sky. Data volumes of entire surveys a decade ago can now be acquired in a single night and real-time analysis is often desired. Thus, modern astronomy requires big data know-how, in particular it demands highly efficient machine learning and image analysis algorithms. But scalability is not the only challenge: Astronomy applications touch several current machine learning research questions, such as learning from biased data and dealing with label and measurement noise. We argue that this makes astronomy a great domain for computer science research, as it pushes the boundaries of data analysis. In the following, we will present this exciting application area for data scientists. We will focus on exemplary results, discuss main challenges, and highlight some recent methodological advancements in machine learning and image analysis triggered by astronomical applications.