The use of Convolutional Neural Networks (CNN) to estimate the galaxy photometric redshift probability distribution by analysing the images in different wavelength bands has been developed in the recent years thanks to the rapid development of the Machine Learning (ML) ecosystem. Authors have set-up CNN architectures and studied their performances and some sources of systematics using standard methods of training and testing to ensure the generalisation power of their models. So far so good, but one piece was missing : does the model generalisation power is well measured? The present article shows clearly that very small image perturbations can fool the model completely and opens the Pandora's box of \textit{adversarial} attack. Among the different techniques and scenarios, we have chosen to use the Fast Sign Gradient one-step Method and its Projected Gradient Descent iterative extension as adversarial generator tool kit. However, as unlikely as it may seem these adversarial samples which fool not only a single model, reveal a weakness both of the model and the classical training. A revisited algorithm is shown and applied by injecting a fraction of adversarial samples during the training phase. Numerical experiments have been conducted using a specific CNN model for illustration although our study could be applied to other models - not only CNN ones - and in other contexts - not only redshift measurements - as it deals with the complexity of the boundary decision surface.

PAON4 is an L-band (1250-1500 MHz) small interferometer operating in transit mode deployed at the Nan\c{c}ay observatory in France, designed as a prototype instrument for Intensity Mapping. It features four 5~meter diameter dishes in a compact triangular configuration, with a total geometric collecting area of $\sim75 \mathrm{m^2}$, and equipped with dual polarization receivers. A total of 36 visibilities are computed from the 8 independent RF signals by the software correlator over the full 250~MHz RF band. The array operates in transit mode, with the dishes pointed toward a fixed declination, while the sky drifts across the instrument. Sky maps for each frequency channel are then reconstructed by combining the time-dependent visibilities from the different baselines observed at different declinations. This paper presents an overview of the PAON4 instrument design and goals, as a prototype for dish arrays to map the Large Scale Structure in radio, using intensity mapping of the atomic hydrogen $21~\mathrm{cm}$ line. We operated PAON4 over several years and use data from observations in different periods to assess the array performance. We present preliminary analysis of a large fraction of this data and discuss crucial issues for this type of instrument, such as the calibration strategy, instrument response stability, and noise behaviour.

Apache Spark is a Big Data framework for working on large distributed datasets. Although widely used in the industry, it remains rather limited in the academic community or often restricted to software engineers. The goal of this paper is to show with practical uses-cases that the technology is mature enough to be used without excessive programming skills by astronomers or cosmologists in order to perform standard analyses over large datasets, as those originating from future galaxy surveys. To demonstrate it, we start from a realistic simulation corresponding to 10 years of LSST data taking (6 billions of galaxies). Then, we design, optimize and benchmark a set of Spark python algorithms in order to perform standard operations as adding photometric redshift errors, measuring the selection function or computing power spectra over tomographic bins. Most of the commands execute on the full 110 GB dataset within tens of seconds and can therefore be performed interactively in order to design full-scale cosmological analyses. A jupyter notebook summarizing the analysis is available at https://github.com/astrolabsoftware/1807.03078.

The Core Cosmology Library (CCL) provides routines to compute basic cosmological observables to a high degree of accuracy, which have been verified with an extensive suite of validation tests. Predictions are provided for many cosmological quantities, including distances, angular power spectra, correlation functions, halo bias and the halo mass function through state-of-the-art modeling prescriptions available in the literature. Fiducial specifications for the expected galaxy distributions for the Large Synoptic Survey Telescope (LSST) are also included, together with the capability of computing redshift distributions for a user-defined photometric redshift model. A rigorous validation procedure, based on comparisons between CCL and independent software packages, allows us to establish a well-defined numerical accuracy for each predicted quantity. As a result, predictions for correlation functions of galaxy clustering, galaxy-galaxy lensing and cosmic shear are demonstrated to be within a fraction of the expected statistical uncertainty of the observables for the models and in the range of scales of interest to LSST. CCL is an open source software package written in C, with a python interface and publicly available at https://github.com/LSSTDESC/CCL.

The Tianlai Pathfinder is designed to demonstrate the feasibility of using a wide field of view radio interferometers to map the density of neutral hydrogen in the Universe after the Epoch of Reionizaton. This approach, called 21~cm intensity-mapping, promises an inexpensive means for surveying the large-scale structure of the cosmos. The Tianlai Pathfinder presently consists of an array of three, 15~m $\times$ 40~m cylinder telescopes and an array of sixteen, 6~m diameter dish antennas located in a radio-quiet part of western China. The two types of arrays were chosen to determine the advantages and disadvantages of each approach. The primary goal of the Pathfinder is to make 3D maps by surveying neutral hydrogen over large areas of the sky %$20,000 {\rm deg}^2$ in two different redshift ranges: first at $1.03 > z > 0.78$ ($700 - 800$~MHz) and later at $0.21 > z > 0.12$ ($1170 - 1270$~MHz). The most significant challenge to $21$~cm intensity-mapping is the removal of strong foreground radiation that dwarfs the cosmological signal. It requires exquisite knowledge of the instrumental response, i.e. calibration. In this paper, we provide an overview of the status of the Pathfinder and discuss the details of some of the analysis that we have carried out to measure the beam function of both arrays. We compare electromagnetic simulations of the arrays to measurements, discuss measurements of the gain and phase stability of the instrument, and provide a brief overview of the data processing pipeline.

In the near future, cosmology will enter the wide and deep galaxy survey area allowing high-precision studies of the large scale structure of the universe in three dimensions. To test cosmological models and determine their parameters accurately, it is natural to confront data with exact theoretical expectations expressed in the observational parameter space (angles and redshift). The data-driven galaxy number count fluctuations on redshift shells, can be used to build correlation functions $C(\theta; z_1, z_2)$ on and between shells which can probe the baryonic acoustic oscillations, the distance-redshift distortions as well as gravitational lensing and other relativistic effects. Transforming the model to the data space usually requires the computation of the angular power spectrum $C_\ell(z_1, z_2)$ but this appears as an artificial and inefficient step plagued by apodization issues. In this article we show that it is not necessary and present a compact expression for $C(\theta; z_1, z_2)$ that includes directly the leading density and redshift space distortions terms from the full linear theory. It can be evaluated using a fast integration method based on Clenshaw-Curtis quadrature and Chebyshev polynomial series. This new method to compute the correlation functions without any Limber approximation, allows us to produce and discuss maps of the correlation function directly in the observable space and is a significant step towards disentangling the data from the tested models.

The statistical distribution of galaxies is a powerful probe to constrain cosmological models and gravity. In particular the matter power spectrum $P(k)$ brings information about the cosmological distance evolution and the galaxy clustering together. However the building of $P(k)$ from galaxy catalogues needs a cosmological model to convert angles on the sky and redshifts into distances, which leads to difficulties when comparing data with predicted $P(k)$ from other cosmological models, and for photometric surveys like LSST. The angular power spectrum $C_\ell(z_1,z_2)$ between two bins located at redshift $z_1$ and $z_2$ contains the same information than the matter power spectrum, is free from any cosmological assumption, but the prediction of $C_\ell(z_1,z_2)$ from $P(k)$ is a costly computation when performed exactly. The Angpow software aims at computing quickly and accurately the auto ($z_1=z_2$) and cross ($z_1 \neq z_2$) angular power spectra between redshift bins. We describe the developed algorithm, based on developments on the Chebyshev polynomial basis and on the Clenshaw-Curtis quadrature method. We validate the results with other codes, and benchmark the performance. Angpow is flexible and can handle any user defined power spectra, transfer functions, and redshift selection windows. The code is fast enough to be embedded inside programs exploring large cosmological parameter spaces through the $C_\ell(z_1,z_2)$ comparison with data. We emphasize that the Limber's approximation, often used to fasten the computation, gives wrong $C_\ell$ values for cross-correlations.

In this paper, we apply our sky map reconstruction method for transit type interferometers to the Tianlai cylinder array. The method is based on the spherical harmonic decomposition, and can be applied to cylindrical array as well as dish arrays and we can compute the instrument response, synthesised beam, transfer function and the noise power spectrum. We consider cylinder arrays with feed spacing larger than half wavelength, and as expected, we find that the arrays with regular spacing have grating lobes which produce spurious images in the reconstructed maps. We show that this problem can be overcome, using arrays with different feed spacing on each cylinder. We present the reconstructed maps, and study the performance in terms of noise power spectrum, transfer function and beams for both regular and irregular feed spacing configurations.

The spherical harmonics $m$-mode decomposition is a powerful sky map reconstruction method suitable for radio interferometers operating in transit mode. It can be applied to various configurations, including dish arrays and cylinders. We describe the computation of the instrument response function, the point spread function (PSF), transfer function, the noise covariance matrix and noise power spectrum. The analysis in this paper is focused on dish arrays operating in transit mode. We show that arrays with regular spacing have more pronounced side lobes as well as structures in their noise power spectrum, compared to arrays with irregular spacing, specially in the north-south direction. A good knowledge of the noise power spectrum $C^{\mathrm{noise}}(\ell)$ is essential for intensity mapping experiments as non uniform $C^{\mathrm{noise}}(\ell)$ is a potential problem for the measurement of the HI power spectrum. Different configurations have been studied to optimise the PAON-4 and Tianlai dish array layouts. We present their expected performance and their sensitivities to the 21-cm emission of the Milky Way and local extragalactic HI clumps

We have observed regions of three galaxy clusters at z$\sim$ [0.06, 0.09] (Abell85, Abell1205, Abell2440), as well as calibration sources with the Nancay radiotelescope (NRT) to search for 21 cm emission and fully characterize the FPGA based BAORadio digital backend. The total observation time of few hours per source have been distributed over few months, from March 2011 to January 2012, due to scheduling constraints of the NRT, which is a transit telescope. Data have been acquired in parallel with the NRT standard correlator (ACRT) back-end, as well as with the BAORadio data acquisition system. The latter enables wide band instantaneous observation of the [1250, 1500]MHz frequency range, as well as the use of powerful RFI mitigation methods thanks to its fine time sampling. A number of questions related to instrument stability, data processing and calibration are discussed. We have obtained the radiometer curves over the integration time range [0.01,10 000] seconds and we show that sensitivities of few mJy over most of the wide frequency band can be reached with the NRT. It is clearly shown that in blind line search, which is the context of HI intensity mapping for Baryon Acoustic Oscillations, the new acquisition system and processing pipeline outperforms the standard one. We report a positive detection of 21 cm emission at 3 sigma-level from galaxies in the outer region of Abell85 at 1352 MHz (14 400 km/s) corresponding to a line strength of 0.8 Jy km/s. We observe also an excess power around 1318 MHz (21 600 km/s), although at lower statistical significance, compatible with emission from Abell1205 galaxies. Detected radio line emissions have been cross matched with optical catalogs and we have derived hydrogen mass estimates.

3D mapping of matter distribution in the universe through the 21 cm radio emission of atomic hydrogen HI is a complementary approach to optical surveys for the study of the Large Scale Structures, in particular for measuring the BAO (Baryon Acoustic Oscillation) scale up to redshifts z < 3, and therefore constraining dark energy parameters. We propose a novel method to map the HI mass distribution in three dimensions in radio, without detecting or identifying individual compact sources. This method would require an instrument with a large instantaneous bandwidth (> 100 MHz) and high sensitivity, while a rather modest angular resolution (~ 10 arcmin) should be sufficient. These requirements can be met by a dense interferometric array or a phased array (FPA) in the focal plane of a large primary reflector, representing a total collecting area of a few thousand square meters with few hundred simultaneous beams covering a 20 to 100 square degrees field of view. We describe the development and qualification of an electronic and data processing system for digital radio interferometry and beam forming suitable for such instruments with several hundred receiver elements.

Large Scale Structures (LSS) in the universe can be traced using the neutral atomic hydrogen HI through its 21cm emission. Such a 3D matter distribution map can be used to test the Cosmological model and to constrain the Dark Energy properties or its equation of state. A novel approach, called intensity mapping can be used to map the HI distribution, using radio interferometers with large instantaneous field of view and waveband. In this paper, we study the sensitivity of different radio interferometer configurations, or multi-beam instruments for the observation of large scale structures and BAO oscillations in 21cm and we discuss the problem of foreground removal. For each configuration, we determine instrument response by computing the (u,v) or Fourier angular frequency plane coverage using visibilities. The (u,v) plane response is the noise power spectrum, hence the instrument sensitivity for LSS P(k) measurement. We describe also a simple foreground subtraction method to separate LSS 21 cm signal from the foreground due to the galactic synchrotron and radio sources emission. We have computed the noise power spectrum for different instrument configuration as well as the extracted LSS power spectrum, after separation of 21cm-LSS signal from the foregrounds. We have also obtained the uncertainties on the Dark Energy parameters for an optimized 21 cm BAO survey. We show that a radio instrument with few hundred simultaneous beams and a collecting area of ~10000 m^2 will be able to detect BAO signal at redshift z ~ 1 and will be competitive with optical surveys.

3D mapping of matter distribution in the universe through the 21 cm radio emission of atomic hydrogen is a complementary approach to optical surveys for the study of the Large Scale Structures, in particular for measuring the BAO (Baryon Acoustic Oscillation) scale up to redshifts z <~ 3 and constrain dark energy. We propose to carry such a survey through a novel method, called intensity mapping, without detecting individual galaxies radio emission. This method requires a wide band instrument, 100 MHz or larger, and multiple beams, while a rather modest angular resolution of 10 arcmin would be sufficient. The instrument would have a few thousand square meters of collecting area and few hundreds of simultaneous beams. These constraints could be fulfilled with a dense array of receivers in interferometric mode, or a phased array at the focal plane of a large antenna.

The feasibility of a next generation neutrino observatory in Europe is being considered within the LAGUNA design study. To accommodate giant neutrino detectors and shield them from cosmic rays, a new very large underground infrastructure is required. Seven potential candidate sites in different parts of Europe and at several distances from CERN are being studied: Boulby (UK), Canfranc (Spain), Fr\'ejus (France/Italy), Pyh\"asalmi (Finland), Polkowice-Sieroszowice (Poland), Slanic (Romania) and Umbria (Italy). The design study aims at the comprehensive and coordinated technical assessment of each site, at a coherent cost estimation, and at a prioritization of the sites within the summer 2010.