Neural networks and the separation of cosmic microwave background and astrophysical signals in sky maps

We implement an independent component analysis (ICA) algorithm to separate signals of different origin in sky maps at several frequencies. Owing to its self-organizing capability, it works without prior assumptions on either the frequency dependence or the angular power spectrum of the various signals; rather, it learns directly from the input data how to identify the statistically independent components, on the assumption that all but, at most, one of the components have non-Gaussian distributions.
We have applied the ICA algorithm to simulated patches of the sky at the four frequencies (30, 44, 70 and 100 GHz) used by the Low Frequency Instrument of the European Space Agency's Planck satellite. Simulations include the cosmic microwave background (CMB), the synchrotron and thermal dust emissions, and extragalactic radio sources. The effects of the angular response functions of the detectors and of instrumental noise have been ignored in this first exploratory study. The ICA algorithm reconstructs the spatial distribution of each component with rms errors of about 1 per cent for the CMB, and 10 per cent for the much weaker Galactic components. Radio sources are almost completely recovered down to a flux limit corresponding to ≃0.7 σ _CMB, where σ _CMB is the rms level of the CMB fluctuations. The signal recovered has equal quality on all scales larger than the pixel size. In addition, we show that for the strongest components (CMB and radio sources) the frequency scaling is recovered with per cent precision. Thus, algorithms of the type presented here appear to be very promising tools for component separation. On the other hand, we have been dealing here with a highly idealized situation. Work to include instrumental noise, the effect of different resolving powers at different frequencies and a more complete and realistic characterization of astrophysical foregrounds is in progress.     

Cosmic microwave background polarisation: foreground contrast and component separation

We evaluate the expected level of foreground contamination to the cosmic microwave background (CMB) polarised radiation, focusing on the diffuse emission from our own Galaxy. In particular, we perform a first attempt to simulate an all sky template of polarised emission from thermal dust. This study indicates that the foreground contamination to CMB B-modes is likely to be relevant on all frequencies, and even at high Galactic latitudes. We review the recent developments in the design of data analysis techniques dedicated to the separation of CMB and foreground emissions in multi-frequency observations, exploiting their statistical independence. We argue that the high quality and detail of the present CMB observations represent an almost ideal statistical dataset where these algorithms can operate with excellent performance. We explicitly show that the recovery of CMB B-modes is possible even if they are well below the foreground level, working at the arcminute resolution at an almost null computational cost. This capability well represents the great potentiality of these new data analysis techniques, which should be seriously taken into account for implementation in present and future CMB observations.     

Estimating the spectral indices of correlated astrophysical foregrounds by a second-order statistical approach

We present the first tests of a new method, the correlated component analysis (CCA) based on second-order statistics, to estimate the mixing matrix, a key ingredient to separate astrophysical foregrounds superimposed to the Cosmic Microwave Background (CMB). In the present application, the mixing matrix is parametrized in terms of the spectral indices of Galactic synchrotron and thermal dust emissions, while the free–free spectral index is prescribed by basic physics, and is thus assumed to be known. We consider simulated observations of the microwave sky with angular resolution and white stationary noise at the nominal levels for the Planck satellite, and realistic foreground emissions, with a position-dependent synchrotron spectral index. We work with two sets of Planck frequency channels: the low-frequency set, from 30 to 143 GHz, complemented with the Haslam 408 MHz map, and the high-frequency set, from 217 to 545 GHz. The concentration of intense free–free emission on the Galactic plane introduces a steep dependence of the spectral index of the global Galactic emission with Galactic latitude, close to the Galactic equator. This feature makes difficult for the CCA to recover the synchrotron spectral index in this region, given the limited angular resolution of Planck , especially at low frequencies. A cut of a narrow strip around the Galactic equator  (| b | < 3°)  , however, allows us to overcome this problem. We show that, once this strip is removed, the CCA allows an effective foreground subtraction, with residual uncertainties inducing a minor contribution to errors on the recovered CMB power spectrum.     

Bayesian foreground analysis with CMB data

The quality of CMB observations has improved dramatically in the last few years, and will continue to do so in the coming decade. Over a wide range of angular scales, the uncertainty due to instrumental noise is now small compared to the cosmic variance. One may claim with some justification that we have entered the era of precision CMB cosmology. However, some caution is still warranted: The errors due to residual foreground contamination in the CMB power spectrum and cosmological parameters remain largely unquantified, and the effect of these errors on important cosmological parameters such as the optical depth τ and spectral index n_s is not obvious. A major goal for current CMB analysis efforts must therefore be to develop methods that allow us to propagate such uncertainties from the raw data through to the final products. Here we review a recently proposed method that may be a first step towards that goal.     

Multidetector multicomponent spectral matching and applications for cosmic microwave background data analysis

We present a new method for analysing multidetector maps containing several astrophysical components. Our method, based on matching the data to a model in the spectral domain, permits us to estimate jointly the spatial power spectra of the components and of the noise, as well as their mixing coefficients. It is of particular relevance for analysis of millimetre-wave maps of cosmic microwave background (CMB) anisotropies.     

Maximum likelihood algorithm for parametric component separation in cosmic microwave background experiments

We discuss an approach to the component separation of microwave, multifrequency sky maps as those typically produced from cosmic microwave background (CMB) anisotropy data sets. The algorithm is based on the two-step, parametric, likelihood-based technique recently elaborated on by Eriksen et al., where the foreground spectral parameters are estimated prior to the actual separation of the components. In contrast with the previous approaches, we accomplish the former task with help of an analytically derived likelihood function for the spectral parameters, which, we show, yields estimates equal to the maximum likelihood values of the full multidimensional data problem. We then use these estimates to perform the second step via the standard, generalized-least-squares-like procedure. We demonstrate that the proposed approach is equivalent to a direct maximization of the full data likelihood, which is recast in a computationally tractable form. We use the corresponding curvature matrices to characterize statistical properties of the recovered parameters. We incorporate in the formalism some of the essential features of the CMB data sets, such as inhomogeneous pixel domain noise, unknown map offsets as well as calibration errors and study their consequences for the separation. We find that the calibration is likely to have a dominant effect on the precision of the spectral parameter determination for a realistic CMB experiment. We apply the algorithm to simulated data and discuss the results. Our focus is on partial sky, total intensity and polarization, CMB experiments such as planned balloon-borne and ground-based efforts, however, the techniques presented here should be also applicable to the full-sky data as for instance, those produced by the Wilkinson Microwave Anisotropy Probe ( WMAP ) satellite and anticipated from the Planck mission.     

Separation of foreground and background signals in single frequency measurements of the CMB polarization

The polarization of the Cosmic Microwave Background (CMB) is a powerful observational tool at hand for modern cosmology. It allows to break the degeneracy of fundamental cosmological parameters one cannot obtain using only anisotropy data and provides new insight into conditions existing in the very early Universe. Many experiments are now in progress whose aim is detecting anisotropy and polarization of the CMB. Measurements of the CMB polarization are however hampered by the presence of polarized foregrounds, above all the synchrotron emission of our Galaxy, whose importance increases as frequency decreases and dominates the polarized diffuse radiation at frequencies below ≃50 GHz. In the past the separation of CMB and synchrotron was made combining observations of the same area of sky at different frequencies. In this paper, we show that the statistical properties of the polarized components of the synchrotron and dust foregrounds are different from the statistical properties of the polarized component of the CMB, therefore one can build a statistical estimator which allows to extract the polarized component of the CMB from single frequency data also when the polarized CMB signal is just a fraction of the total polarized signal. Our estimator improves the signal/noise ratio for the polarized component of the CMB and reduces from ≃50 to ≃20 GHz, the frequency above which the polarized component of the CMB can be extracted from single frequency maps of the diffuse radiation.     

Foreground removal from Planck Sky Model temperature maps using a MLP neural network

Unfortunately, the Cosmic Microwave Background (CMB) radiation is contaminated by emission originating in the Milky Way (synchrotron, free‐free and dust emission). Since the cosmological information is statistically in nature, it is essential to remove this foreground emission and leave the CMB with no systematic errors. To demonstrate the feasibility of a simple multilayer perceptron (MLP) neural network for extracting the CMB temperature signal, we have analyzed a specific data set, namely the Planck Sky Model maps, developed for evaluation of different component separation methods before including them in the Planck data analysis pipeline. It is found that a MLP neural network can provide a CMB map of about 80 % of the sky to a very high degree uncorrelated with the foreground components. Also the derived power spectrum shows little evidence for systematic errors (© 2009 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Foreground removal from CMB temperature maps using an MLP neural network

One of the main obstacles for extracting the Cosmic Microwave Background (CMB) signal from observations in the mm-submm range is the foreground contamination by emission from Galactic components: mainly synchrotron, free-free and thermal dust emission. Due to the statistical nature of the intrinsic CMB signal it is essential to minimize the systematic errors in the CMB temperature determinations. Following the available knowledge of the spectral behavior of the Galactic foregrounds simple power law-like spectra have been assumed. The feasibility of using a simple neural network for extracting the CMB temperature signal from the combined signal CMB and the foregrounds has been investigated. As a specific example, we have analysed simulated data, as expected from the ESA Planck CMB mission. A simple multilayer perceptron neural network with 2 hidden layers can provide temperature estimates over more than 80 per cent of the sky that are to a high degree uncorrelated with the foreground signals. A single network will be able to cover the dynamic range of the Planck noise level over the entire sky.     

2D/3D Dust Continuum Radiative Transfer Codes to Analyze and Predict VLTI Observations

Radiative Transfer (RT) codes with image capability are a fundamental tool for preparing interferometric observations and for interpreting visibility data. In view of the upcoming VLTI facilities, we present the first comparison of images/visibilities coming from two 3D codes that use completely different techniques to solve the problem of self-consistent continuum RT. In addition, we focus on the astrophysical case of a disk distorted by tidal interaction with by-passing stars or internal planets and investigate for which parameters the distortion can be best detected in the mid-infrared using the mid-infrared interferometric device MIDI. This revised version was published online in July 2006 with corrections to the Cover Date.     

A preferred-direction statistic for sky maps

Large patterns could exist on the microwave sky as a result of various non-standard possibilities for the large-scale Universe – rotation or shear, non-trivial topology, and single topological defects are specific examples. All-sky (or nearly all-sky) CMB data sets allow us, uniquely, to constrain such exotica, and it is therefore worthwhile to explore a wide range of statistical tests. We describe one such statistic here, which is based on determining gradients and is useful for assessing the level of 'preferred directionality' or 'stripiness' in the map. This method is more general than other techniques for picking out specific patterns on the sky, and it also has the advantage of being easily calculable for the mega-pixel maps which will soon be available. For the purposes of illustration, we apply this statistic to the four-year COBE DMR data. For future CMB maps, we expect this to be a useful statistical test of the large-scale structure of the Universe. In principle, the same statistic could also be applied to sky maps at other wavelengths, to CMB polarization maps, and to catalogues of discrete objects. It may also be useful as a means of checking for residual directionality (e.g. from Galactic or ecliptic signals) in maps.     

Map-making methods for cosmic microwave background experiments

The map-making step of cosmic microwave background (CMB) data analysis involves linear inversion problems that cannot be performed by a brute-force approach for the large time-lines of today. In this paper we present optimal vector-only map-making methods, which are an iterative COBE method, a Wiener direct filter and a Wiener iterative method. We apply these methods on diverse simulated data, and we show that they produce very well restored maps, by removing nearly completely the correlated noise that appears as intense stripes on the simply pixel-averaged maps. The COBE iterative method can be applied to any signals, assuming the stationarity of the noise in the time-line. The Wiener methods assume both the stationarity of the noise and the sky, which is the case for CMB-only data. We apply the methods to Galactic signals too, and test them on balloon-borne experiment strategies and on a satellite whole-sky survey.     

Decay of the vacuum energy into cosmic microwave background photons

We examine the possibility of the decay of the vacuum energy into a homogeneous distribution of a thermalized cosmic microwave background (CMB), which is characteristic of an adiabatic vacuum energy decay into photons. It is shown that observations of the primordial density fluctuation spectrum, obtained from CMB and galaxy distribution data, restrict the possible decay rate. When photon creation due to an adiabatic vacuum energy decay takes place, the standard linear temperature dependence   T ( z ) = T ₀(1 + z )  is modified, where T ₀ is the present CMB temperature, and can be parametrized by a modified CMB temperature dependence     . From the observed CMB and galaxy distribution data, a strong limit on the maximum value of the decay rate is obtained by placing a maximum value  β_max≃ 3.4 × 10⁻³  on the β parameter.     

The inverse problem for pulsating neutron stars: a 'fingerprint analysis' for the supranuclear equation of state

We study the problem of detecting, and inferring astrophysical information from, gravitational waves from a pulsating neutron star. We show that the fluid f and p modes, as well as the gravitational-wave w modes, may be detectable from sources in our own Galaxy, and investigate how accurately the frequencies and damping rates of these modes can be inferred from a noisy gravitational-wave data stream. Based on the conclusions of this discussion we propose a strategy for revealing the supranuclear equation of state using the neutron star fingerprints: the observed frequencies of an f and a p mode. We also discuss how well the source can be located in the sky using observations with several detectors.     

Extended inverse-Compton emission from distant, powerful radio galaxies

We present Chandra observations of two relatively high redshift FR II radio galaxies, 3C 432 and 3C 191 (   z = 1.785  and 1.956, respectively), both of which show extended X-ray emission along the axis of the radio jet or lobe. This X-ray emission is most likely to be due to inverse-Compton scattering of cosmic microwave background (CMB) photons. Under this assumption, we estimate the minimum energy contained in the particles responsible. This can be extrapolated to determine a rough estimate of the total energy. We also present new, deep radio observations of 3C 294, which confirm some association between radio and X-ray emission along the north-east–south-west radio axis and also that radio emission is not detected over the rest of the extent of the diffuse X-ray emission. This together with the offset between the peaks of the X-ray and radio emissions may indicate that the jet axis in this source is precessing.     

Bayesian joint analysis of cluster weak lensing and Sunyaev–Zel'dovich effect data

As the quality of the available galaxy cluster data improves, the models fitted to these data might be expected to become increasingly complex. Here we present the Bayesian approach to the problem of cluster data modelling: starting from simple, physically motivated parametrized functions to describe the cluster's gas density, gravitational potential and temperature, we explore the high-dimensional parameter spaces with a Markov-Chain Monte Carlo sampler, and compute the Bayesian evidence in order to make probabilistic statements about the models tested. In this way sufficiently good data will enable the models to be distinguished, enhancing our astrophysical understanding; in any case the models may be marginalized over in the correct way when estimating global, perhaps cosmological, parameters. In this work we apply this methodology to two sets of simulated interferometric Sunyaev–Zel'dovich effect and gravitational weak lensing data, corresponding to current and next-generation telescopes. We calculate the expected precision on the measurement of the cluster gas fraction from such experiments, and investigate the effect of the primordial cosmic microwave background (CMB) fluctuations on their accuracy. We find that data from instruments such as the Arcminute Microkelvin Imager (AMI), when combined with wide-field ground-based weak lensing data, should allow both cluster model selection and estimation of gas fractions to a precision of better than 30 per cent for a given cluster.     

Effects of destriping errors on estimates of the CMB power spectrum

Destriping methods for constructing maps of the cosmic microwave background (CMB) anisotropies have been investigated extensively in the literature. However, their error properties have been studied in less detail. Here we present an analysis of the effects of destriping errors on CMB power spectrum estimates for Planck -like scanning strategies. Analytic formulae are derived for certain simple scanning geometries that can be rescaled to account for different detector noise. Assuming Planck -like low-frequency noise, the noise power spectrum is accurately white at high multipoles  (ℓ≳ 50)  . Destriping errors, though dominant at lower multipoles, are small in comparison to the cosmic variance. These results show that simple destriping map-making methods should be perfectly adequate for the analysis of Planck data and support the arguments given in an earlier paper in favour of applying a fast hybrid power spectrum estimator to CMB data with realistic '1/ f ' noise.     

Cross-correlation between WMAP and 2MASS: non-Gaussianity induced by the SZ effect

We study the non-Gaussianity induced by the Sunyaev–Zel'dovich (SZ) effect in cosmic microwave background (CMB) fluctuation maps. If a CMB map is contaminated by the SZ effect of galaxies or galaxy clusters, the CMB maps should have similar non-Gaussian features to the galaxy and cluster fields. Using the WMAP data and 2MASS galaxy catalogue, we show that the non-Gaussianity of the 2MASS galaxies is imprinted on WMAP maps. The signature of non-Gaussianity can be seen with the fourth-order cross-correlation between the wavelet variables of the WMAP maps and 2MASS clusters. The intensity of the fourth-order non-Gaussian features is found to be consistent with the contamination of the SZ effect of 2MASS galaxies. We also show that this non-Gaussianity can not be seen by the high-order autocorrelation of the WMAP . This is because the SZ signals in the autocorrelations of the WMAP data generally are weaker than the WMAP –2MASS cross-correlations by a factor f ², which is the ratio between the powers of the SZ-effect map and the CMB fluctuations on the scale considered. Therefore, the ratio of high-order autocorrelations of CMB maps to cross-correlations of the CMB maps and galaxy field would be effective to constrain the powers of the SZ effect on various scales.     

Imprints of relic gravitational waves on pulsar timing

Relic gravitational waves(RGWs), a background originating during inflation, would leave imprints on pulsar timing residuals. This makes RGWs an important source for detection of RGWs using the method of pulsar timing. In this paper, we discuss the effects of RGWs on single pulsar timing, and quantitatively analyze the timing residuals caused by RGWs with different model parameters. In principle, if the RGWs are strong enough today, they can be detected by timing a single millisecond pulsar with high precision after the intrinsic red noises in pulsar timing residuals are understood, even though simultaneously observing multiple millisecond pulsars is a more powerful technique for extracting gravitational wave signals. We correct the normalization of RGWs using observations of the cosmic microwave background(CMB), which leads to the amplitudes of RGWs being reduced by two orders of magnitude or so compared to our previous works. We obtained new constraints on RGWs using recent observations from the Parkes Pulsar Timing Array, employing the tensor-to-scalar ratio r = 0.2 due to the tensor-type polarization observations of CMB by BICEP2 as a reference value, even though its reliability has been brought into question. Moreover, the constraints on RGWs from CMB and Big Bang nucleosynthesis will also be discussed for comparison.     

Observations of the Sunyaev–Zel''dovich effect at high angular resolution towards the galaxy clusters A665, A2163 and CL0016+16

We report on the first observation of the Sunyaev–Zel'dovich (SZ) effect, a distortion of the Cosmic Microwave Background radiation (CMB) by hot electrons in clusters of galaxies, with the Diabolo experiment at the IRAM 30 m telescope. Diabolo is a dual-channel 0.1 K bolometer photometer dedicated to the observation of CMB anisotropies at 2.1 and 1.2 mm. A significant brightness decrement in the 2.1 mm channel is detected in the direction of three clusters (Abell 665, Abell 2163 and CL0016+16). With a 30 arcsec beam and 3 arcmin beamthrow, this is the highest angular resolution observation to date of the SZ effect. Interleaving integrations on targets and on nearby blank fields have been performed in order to check and correct for systematic effects. Gas masses can be directly inferred from these observations.