Upcoming large astronomical surveys are expected to capture an unprecedented number of strong gravitational lensing systems in the Universe. Deep learning is emerging as a promising practical tool in detection and quantification of these galaxy-scale image distortions. However, absence of large quantities of representative data from current astronomical surveys requires development of robust forward modeling of synthetic lensing images. Using a realistic and unbiased sample of the strong lenses created by using state-of-the-art extragalactic catalogs, we train a modular deep learning pipeline for uncertainty-quantified detection and modeling with intermediate image processing components for denoising and deblending the lensing systems. We demonstrate a higher degree of interpretability and controlled systematics due to domain-specific task modules that are trained with different stages of synthetic image generation. For lens detection and modeling, we obtain semantically meaningful latent spaces that separate classes and provide uncertainty estimates that explain the misclassified images and provide uncertainty bounds on the lens parameters. In addition, we obtain an improved performance---lens detection (classification) improved from 82% with the baseline to 94%, while the lens modeling (regression) accuracy improved by 25% over the baseline model.

Large pristine samples of red clump stars are highly sought after given that they are standard candles and give precise distances even at large distances. However, it is difficult to cleanly select red clumps stars because they can have the same T$_{\mathrm{eff}}$ and log $g$ as red giant branch stars. Recently, it was shown that the asteroseismic parameters, $\rm{\Delta}$P and $\rm{\Delta\nu}$, which are used to accurately select red clump stars, can be derived from spectra using the change in the surface carbon to nitrogen ratio ([C/N]) caused by mixing during the red giant branch. This change in [C/N] can also impact the spectral energy distribution. In this study, we predict the $\rm{\Delta}$P, $\rm{\Delta\nu}$, T$_{\mathrm{eff}}$ and log $g$ using 2MASS, AllWISE, \gaia, and Pan-STARRS data in order to select a clean sample of red clump stars. We achieve a contamination rate of $\sim$20\%, equivalent to what is achieved when selecting from T$_{\mathrm{eff}}$ and log $g$ derived from low resolution spectra. Finally, we present two red clump samples. One sample has a contamination rate of $\sim$ 20\% and $\sim$ 405,000 red clump stars. The other has a contamination of $\sim$ 33\% and $\sim$ 2.6 million red clump stars which includes $\sim$ 75,000 stars at distances $>$ 10 kpc. For |b|>30 degrees we find $\sim$ 15,000 stars with contamination rate of $\sim$ 9\%. The scientific potential of this catalog for studying the structure and formation history of the Galaxy is vast given that it includes millions of precise distances to stars in the inner bulge and distant halo where astrometric distances are imprecise.

We describe the observations and resultant galaxy cluster catalog from the 2770 deg$^2$ SPTpol Extended Cluster Survey (SPT-ECS). Clusters are identified via the Sunyaev-Zel'dovich (SZ) effect, and confirmed with a combination of archival and targeted follow-up data, making particular use of data from the Dark Energy Survey (DES). With incomplete followup we have confirmed as clusters 244 of 266 candidates at a detection significance $\xi \ge 5$ and an additional 204 systems at $4<\xi<5$. The confirmed sample has a median mass of $M_{500c} \sim {4.4 \times 10^{14} M_\odot h_{70}^{-1}}$, a median redshift of $z=0.49$, and we have identified 44 strong gravitational lenses in the sample thus far. Radio data are used to characterize contamination to the SZ signal; the median contamination for confirmed clusters is predicted to be $\sim$1% of the SZ signal at the $\xi>4$ threshold, and $<4\%$ of clusters have a predicted contamination $>10\% $ of their measured SZ flux. We associate SZ-selected clusters, from both SPT-ECS and the SPT-SZ survey, with clusters from the DES redMaPPer sample, and find an offset distribution between the SZ center and central galaxy in general agreement with previous work, though with a larger fraction of clusters with significant offsets. Adopting a fixed Planck-like cosmology, we measure the optical richness-to-SZ-mass ($\lambda-M$) relation and find it to be 28% shallower than that from a weak-lensing analysis of the DES data---a difference significant at the 4 $\sigma$ level---with the relations intersecting at $\lambda=60$ . The SPT-ECS cluster sample will be particularly useful for studying the evolution of massive clusters and, in combination with DES lensing observations and the SPT-SZ cluster sample, will be an important component of future cosmological analyses.

Matter density is formally infinite at the location of caustic surfaces, where dark matter sheet folds in phase-space. The caustics separate regions with different number of streams and the volume elements change the parity by turning inside out when passing through the caustic stage. Being measure-zero structures, identification of caustics via matter density fields is usually restricted to fine-grained simulations. Here we employ a generic algorithm to identify caustics directly using the triangulation of the Lagrangian sub-manifold x(q,t) obtained in N-body simulations. In our approach the caustic surfaces are approximated by a set of triangles whose vertices are particles of the simulation. The major obstacle we encountered was insufficient sampling of small scale perturbations. We overcame it by a brute force approach. We continued to raise the scale of the cutoff in the initial power spectrum until obtained the reliable resolution of the caustics shells up to seven layers. Although quite modest, our result is the first reliable direct construction of caustic surfaces in N-body simulation. It reveals a number of unexpected geometrical features. In particular shapes of some of them are contrastingly different from the known shapes of the caustics formed in the Zeldovich approximation.

Mysterious dark matter constitutes about 85% of all mass in the Universe. Clustering of dark matter plays the dominant role in the formation of all observed structures on scales from a fraction to a few hundreds of Mega-parsecs. Galaxies play a role of lights illuminating these structures so they can be observed. The observations in the last several decades have unveiled opulent geometry of these structures currently known as the cosmic web. Haloes are the highest concentrations of dark matter and host luminous galaxies. Currently the most accurate modeling of dark matter haloes is achieved in cosmological N-body simulations. Identifying the haloes from the distribution of particles in N-body simulations is one of the problems attracting both considerable interest and efforts. We propose a novel framework for detecting potential dark matter haloes using the field unique for dark matter -- multistream field. The multistream field emerges at the nonlinear stage of the growth of perturbations because the dark matter is collisionless. Counting the number of velocity streams in gravitational collapses supplements our knowledge of spatial clustering. We assume that the virialized haloes have convex boundaries. Closed and convex regions of the multistream field are hence isolated by imposing a positivity condition on all three eigenvalues of the Hessian estimated on the smoothed multistream field. In a single-scale analysis of high multistream field resolution and low softening length, the halo substructures with local multistream maxima are isolated as individual halo sites.

The cosmic web is one of the most striking features of the distribution of galaxies and dark matter on the largest scales in the Universe. It is composed of dense regions packed full of galaxies, long filamentary bridges, flattened sheets and vast low density voids. The study of the cosmic web has focused primarily on the identification of such features, and on understanding the environmental effects on galaxy formation and halo assembly. As such, a variety of different methods have been devised to classify the cosmic web -- depending on the data at hand, be it numerical simulations, large sky surveys or other. In this paper we bring twelve of these methods together and apply them to the same data set in order to understand how they compare. In general these cosmic web classifiers have been designed with different cosmological goals in mind, and to study different questions. Therefore one would not {\it a priori} expect agreement between different techniques however, many of these methods do converge on the identification of specific features. In this paper we study the agreements and disparities of the different methods. For example, each method finds that knots inhabit higher density regions than filaments, etc. and that voids have the lowest densities. For a given web environment, we find substantial overlap in the density range assigned by each web classification scheme. We also compare classifications on a halo-by-halo basis; for example, we find that 9 of 12 methods classify around a third of group-mass haloes (i.e. $M_{\rm halo}\sim10^{13.5}h^{-1}M_{\odot}$) as being in filaments. Lastly, so that any future cosmic web classification scheme can be compared to the 12 methods used here, we have made all the data used in this paper public.

Topological connections in the single-streaming voids and multistreaming filaments and walls reveal a cosmic web structure different from traditional mass density fields. A single void structure not only percolates the multistream field in all the directions, but also occupies over 99 per cent of all the single-streaming regions. Sub-grid analyses on scales smaller than simulation resolution reveal tiny pockets of voids that are isolated by membranes of the structure. For the multistreaming excursion sets, the percolating structure is significantly thinner than the filaments in over-density excursion approach. Hessian eigenvalues of the multistream field are used as local geometrical indicators of dark matter structures. Single-streaming regions have most of the zero eigenvalues. Parameter-free conditions on the eigenvalues in the multistream region may be used to delineate primitive geometries with concavities corresponding to filaments, walls and haloes.

We report the results of the first study of the multi-stream environment of dark matter haloes in cosmological N-body simulations in the LCDM cosmology. The full dynamical state of dark matter can be described as a three-dimensional sub-manifold in six dimensional phase space - the dark matter sheet. In our study we use a Lagrangian sub-manifold x = x(q,t) (where x and q are co-moving Eulerian and Lagrangian coordinates respectively), which is dynamically equivalent to the dark matter sheet but is more convenient for numerical analysis. Our major results can be summarized as follows. At the resolution of the simulation i.e. without additional smoothing, the cosmic web represents a hierarchical structure: each halo is embedded in the filamentary framework of the web predominantly at the filament crossings, and each filament is embedded in the wall like fabric of the web at the wall crossings. Locally, each halo or sub-halo is a peak in the number of streams field. The number of streams in the neighbouring filaments is higher than in the neighbouring walls. The walls are regions where number of streams is equal to three or a few. Voids are uniquely defined by the local condition requiring to be a single-stream flow region. The shells of streams around haloes are quite thin and the closest void region is typically within one and a half FOF radius from the center of the halo.