اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدDeep Learning at Scale for the Construction of Galaxy Catalogs in the Dark Energy Survey
228
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
The scale of ongoing and future electromagnetic surveys pose formidable challenges to classify astronomical objects. Pioneering efforts on this front include citizen science campaigns adopted by the Sloan Digital Sky Survey (SDSS). SDSS datasets have been recently used to train neural network models to classify galaxies in the Dark Energy Survey (DES) that overlap the footprint of both surveys. Herein, we demonstrate that knowledge from deep learning algorithms, pre-trained with real-object images, can be transferred to classify galaxies that overlap both SDSS and DES surveys, achieving state-of-the-art accuracy $gtrsim99.6%$. We demonstrate that this process can be completed within just eight minutes using distributed training. While this represents a significant step towards the classification of DES galaxies that overlap previous surveys, we need to initiate the characterization of unlabelled DES galaxies in new regions of parameter space. To accelerate this program, we use our neural network classifier to label over ten thousand unlabelled DES galaxies, which do not overlap previous surveys. Furthermore, we use our neural network model as a feature extractor for unsupervised clustering and find that unlabeled DES images can be grouped together in two distinct galaxy classes based on their morphology, which provides a heuristic check that the learning is successfully transferred to the classification of unlabelled DES images. We conclude by showing that these newly labeled datasets can be combined with unsupervised recursive training to create large-scale DES galaxy catalogs in preparation for the Large Synoptic Survey Telescope era.
قيم البحث
اقرأ أيضاً
The calibration and validation of scientific analysis in simulations is a fundamental tool to ensure unbiased and robust results in observational cosmology. In particular, mock galaxy catalogs are a crucial resource to achieve these goals in the meas
urement of Baryon Acoustic Oscillations (BAO) in the clustering of galaxies. Here we present a set of 1952 galaxy mock catalogs designed to mimic the Dark Energy Survey (DES) Year 3 BAO sample over its full photometric redshift range $0.6 < z_{rm photo} < 1.1$. The mocks are based upon 488 ICE-COLA fast $N$-body simulations of full-sky light-cones and are created by populating halos with galaxies, using a hybrid Halo Occupation Distribution - Halo Abundance Matching model. This model has 10 free parameters, which are determined, for the first time, using an automatic likelihood minimization procedure. We also introduce a novel technique to assign photometric redshift for simulated galaxies, following a two-dimensional probability distribution with VIMOS Public Extragalactic Redshift Survey (VIPERS) data. The calibration was designed to match the observed abundance of galaxies as a function of photometric redshift, the distribution of photometric redshift errors, and the clustering amplitude on scales smaller than those used for BAO measurements. An exhaustive analysis is done to ensure that the mocks reproduce the input properties. Finally, mocks are tested by comparing the angular correlation function $w(theta)$, angular power spectrum $C_ell$ and projected clustering $xi_p(r_perp)$ to theoretical predictions and data. The success in reproducing accurately the photometric redshift uncertainties and the galaxy clustering as a function of redshift render this mock creation pipeline as a benchmark for future analyses of photometric galaxy surveys.
Compact binary systems emit gravitational radiation which is potentially detectable by current Earth bound detectors. Extracting these signals from the instruments background noise is a complex problem and the computational cost of most current searc
hes depends on the complexity of the source model. Deep learning may be capable of finding signals where current algorithms hit computational limits. Here we restrict our analysis to signals from non-spinning binary black holes and systematically test different strategies by which training data is presented to the networks. To assess the impact of the training strategies, we re-analyze the first published networks and directly compare them to an equivalent matched-filter search. We find that the deep learning algorithms can generalize low signal-to-noise ratio (SNR) signals to high SNR ones but not vice versa. As such, it is not beneficial to provide high SNR signals during training, and fastest convergence is achieved when low SNR samples are provided early on. During testing we found that the networks are sometimes unable to recover any signals when a false alarm probability $<10^{-3}$ is required. We resolve this restriction by applying a modification we call unbounded Softmax replacement (USR) after training. With this alteration we find that the machine learning search retains $geq 97.5%$ of the sensitivity of the matched-filter search down to a false-alarm rate of 1 per month.
We present a suite of 18 synthetic sky catalogs designed to support science analysis of galaxies in the Dark Energy Survey Year 1 (DES Y1) data. For each catalog, we use a computationally efficient empirical approach, ADDGALS, to embed galaxies withi
n light-cone outputs of three dark matter simulations that resolve halos with masses above ~5x10^12 h^-1 m_sun at z <= 0.32 and 10^13 h^-1 m_sun at z~2. The embedding method is tuned to match the observed evolution of galaxy counts at different luminosities as well as the spatial clustering of the galaxy population. Galaxies are lensed by matter along the line of sight --- including magnification, shear, and multiple images --- using CALCLENS, an algorithm that calculates shear with 0.42 arcmin resolution at galaxy positions in the full catalog. The catalogs presented here, each with the same LCDM cosmology (denoted Buzzard), contain on average 820 million galaxies over an area of 1120 square degrees with positions, magnitudes, shapes, photometric errors, and photometric redshift estimates. We show that the weak-lensing shear catalog, redMaGiC galaxy catalogs and redMaPPer cluster catalogs provide plausible realizations of the same catalogs in the DES Y1 data by comparing their magnitude, color and redshift distributions, angular clustering, and mass-observable relations, making them useful for testing analyses that use these samples. We make public the galaxy samples appropriate for the DES Y1 data, as well as the data vectors used for cosmology analyses on these simulations.
We perform a comparison of different approaches to star-galaxy classification using the broad-band photometric data from Year 1 of the Dark Energy Survey. This is done by performing a wide range of tests with and without external `truth information,
which can be ported to other similar datasets. We make a broad evaluation of the performance of the classifiers in two science cases with DES data that are most affected by this systematic effect: large-scale structure and Milky Way studies. In general, even though the default morphological classifiers used for DES Y1 cosmology studies are sufficient to maintain a low level of systematic contamination from stellar mis-classification, contamination can be reduced to the O(1%) level by using multi-epoch and infrared information from external datasets. For Milky Way studies the stellar sample can be augmented by ~20% for a given flux limit. Reference catalogs used in this work will be made available upon publication.
Gravitational waves from the coalescence of compact-binary sources are now routinely observed by Earth bound detectors. The most sensitive search algorithms convolve many different pre-calculated gravitational waveforms with the detector data and loo
k for coincident matches between different detectors. Machine learning is being explored as an alternative approach to building a search algorithm that has the prospect to reduce computational costs and target more complex signals. In this work we construct a two-detector search for gravitational waves from binary black hole mergers using neural networks trained on non-spinning binary black hole data from a single detector. The network is applied to the data from both observatories independently and we check for events coincident in time between the two. This enables the efficient analysis of large quantities of background data by time-shifting the independent detector data. We find that while for a single detector the network retains $91.5%$ of the sensitivity matched filtering can achieve, this number drops to $83.9%$ for two observatories. To enable the network to check for signal consistency in the detectors, we then construct a set of simple networks that operate directly on data from both detectors. We find that none of these simple two-detector networks are capable of improving the sensitivity over applying networks individually to the data from the detectors and searching for time coincidences.
الأسئلة المقترحة
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
