-
PUREPath: A Deep Latent Variational Model for Estimating CMB Posterior over Large Angular Scales of the Sky
Authors:
Vipin Sudevan,
Pisin Chen
Abstract:
We present a comprehensive neural architecture, the PUREPath, which leverages a nested Probabilistic multi-modal U- Net framework, augmented by the inclusion of probabilistic ResNet blocks in the Expanding Pathway of the decoders, to estimate the posterior density of the Cosmic Microwave Background (CMB) signal conditioned on the observed CMB data and the training dataset. By seamlessly integratin…
▽ More
We present a comprehensive neural architecture, the PUREPath, which leverages a nested Probabilistic multi-modal U- Net framework, augmented by the inclusion of probabilistic ResNet blocks in the Expanding Pathway of the decoders, to estimate the posterior density of the Cosmic Microwave Background (CMB) signal conditioned on the observed CMB data and the training dataset. By seamlessly integrating Bayesian statistics and variational methods our model effectively minimizes foreground contamination in the observed CMB maps. The model is trained using foreground and noise contaminated CMB temperature maps simulated at Planck LFI and HFI frequency channels 30 - 353 GHz using publicly available Code for Anisotropies in the Microwave Background (CAMB) and Python Sky Model (PySM) packages. During training, our model transforms initial prior distribution on the model parameters to posterior distributions based on the training data. From the joint full posterior of the model parameters, during inference, a predicitve CMB posterior and summary statistics such as the predictive mean, variance etc of the cleaned CMB map is estimated. The predictive standard deviation map provides a direct and interpretable measure of uncertainty per pixel in the predicted mean CMB map. The cleaned CMB map along with the error estimates can be used for more accurate measurements of cosmological parameters and other cosmological analyses.
△ Less
Submitted 28 June, 2024; v1 submitted 27 June, 2024;
originally announced June 2024.
-
A Partial-Sky Gibbs ILC Approach for the Estimation of CMB Posterior over Large Angular Scales of the Sky
Authors:
Vipin Sudevan,
Ujjal Purkayastha,
Rajib Saha
Abstract:
In this article we present a formalism to incorporate the partial-sky maps to the Gibbs ILC algorithm to estimate the joint posterior density of the Cosmic Microwave Background (CMB) signal and the theoretical CMB angular power spectrum given the observed CMB maps. In order to generate the partial-sky maps we mask all the observed CMB maps provided by WMAP and Planck satellite mission using a Gaus…
▽ More
In this article we present a formalism to incorporate the partial-sky maps to the Gibbs ILC algorithm to estimate the joint posterior density of the Cosmic Microwave Background (CMB) signal and the theoretical CMB angular power spectrum given the observed CMB maps. In order to generate the partial-sky maps we mask all the observed CMB maps provided by WMAP and Planck satellite mission using a Gaussian smoothed mask formed on the basis of thermal dust emissions in Planck 353 GHz map. The central galactic region from all the input maps is removed after the application of the smoothed mask. While implementing the Gibbs ILC method on the partial-sky maps, we convert the partial-sky cleaned angular power spectrum to the full-sky angular power spectrum using the mode-mode coupling matrix estimated from the smoothed mask. The main products of our analysis are partial-sky cleaned best-fit CMB map and an estimate of the underlying full-sky theoretical CMB angular power spectrum along with their error estimates. We validate the methodology by performing detailed Monte Carlo simulations after using realistic models of foregrounds and detector noise consistent with the WMAP and Planck frequency channels used in our analysis. We can estimate the posterior density and full-sky theoretical CMB angular power spectrum, without any need to explicitly model the foreground components, from partial-sky maps using our method. Another important feature of this method is that the power spectrum results along with the error estimates can be directly used for cosmological parameter estimations.
△ Less
Submitted 15 October, 2021;
originally announced October 2021.
-
A Foreground Model Independent Estimation of Joint Posterior of CMB E mode Polarization over Large Angular Scales
Authors:
Ujjal Purkayastha,
Vipin Sudevan,
Rajib Saha
Abstract:
Ever since Cosmic Microwave Background (CMB) signal is being measured by various satellites based observations with increasing experimental accuracies there has been a parallel increase in the demand for a CMB reconstruction technique which can provide accurate estimates of CMB signal and the theoretical angular power spectrum along with reliable statistical error estimates associated with them. I…
▽ More
Ever since Cosmic Microwave Background (CMB) signal is being measured by various satellites based observations with increasing experimental accuracies there has been a parallel increase in the demand for a CMB reconstruction technique which can provide accurate estimates of CMB signal and the theoretical angular power spectrum along with reliable statistical error estimates associated with them. In this work, we estimate the joint posterior of CMB E mode signal (S) and corresponding theoretical angular power spectrum (C^E_l) over large angular scales given the simulated polarization observations of future generation COrE satellite mission. To generate samples from the joint distribution we employ the ILC technique with prior information of CMB E mode covariance matrix augmented by a Gibbs sampling technique. We estimate the marginalized densities of S and C^E_l using the samples from full-posterior. The best fit cleaned E mode map and the corresponding angular power spectrum agree well with the input E mode map and the sky power spectrum implying accurate reconstruction using COrE like observations. Using the samples C^E_l of all Gibbs chains we estimate the likelihood function P(C^E_l|D) of any arbitrary C^E_l given simulated observed maps (D) of COrE mission following Blackwell-Rao estimator. The likelihood function can be seamlessly integrated to the cosmological parameter estimation method. Apart from producing an accurate estimate of E mode signal over large angular scales our method also builds a connection between the component reconstruction and reliable cosmological parameter estimation using CMB E mode observations over large angular scales. The entire method does not assume any explicit models for E mode foreground components in order to remove them, which is an attractive property since foreground modelling uncertainty does not pose as a challenge in this case.
△ Less
Submitted 18 April, 2020;
originally announced April 2020.
-
An Accurate Reconstruction of CMB E Mode Signal over Large Angular Scales using Prior Information of CMB Covariance Matrix in ILC Algorithm
Authors:
Ujjal Purkayastha,
Vipin Sudevan,
Rajib Saha
Abstract:
In the recent years, the internal-linear-combination (ILC) method was investigated extensively in the context of reconstruction of Cosmic Microwave Background (CMB) temperature anisotropy signal using observations obtained by WMAP and Planck satellite missions. In this article, we, for the first time, apply the ILC method to reconstruct the large scale CMB E mode polarization signal, which serves…
▽ More
In the recent years, the internal-linear-combination (ILC) method was investigated extensively in the context of reconstruction of Cosmic Microwave Background (CMB) temperature anisotropy signal using observations obtained by WMAP and Planck satellite missions. In this article, we, for the first time, apply the ILC method to reconstruct the large scale CMB E mode polarization signal, which serves as the unique probe of ionization history of the Universe, using simulated observations of 15 frequency CMB polarization maps of future generation COrE satellite mission. We find that usual ILC cleaned E mode map is highly erroneous due to presence of a chance-correlation between CMB and astrophysical foreground components in the empirical covariance matrix which is used to estimate the weight factors. The cleaned angular power spectrum for E mode is strongly biased and erroneous due to these chance correlation factors. In order to address the issues of bias and errors we extend and improve the usual ILC method for CMB E mode reconstruction by incorporating prior information of theoretical E mode angular power spectrum while estimating the weights for linear combination of input maps. Using the E mode covariance matrix effectively suppresses the CMB-foreground chance correlation power leading to an accurate reconstruction of cleaned CMB E mode map and its angular power spectrum. We provide a comparative study of the performance of the usual ILC and the new method over large angular scales of the sky and show that the later produces significantly statistically improved results than the former. The new E mode CMB angular power spectrum contains neither any significant negative bias at the low multipoles nor any positive foreground bias at relatively higher mutlipoles. The error estimates of the cleaned spectrum agree very well with the cosmic variance induced error.
△ Less
Submitted 30 March, 2020;
originally announced March 2020.
-
On the Impact of Random Residual Calibration Error on the Gibbs ILC Estimates over Large Angular Scales
Authors:
Vipin Sudevan,
Rajib Saha
Abstract:
Residual error in calibration coefficients corresponding to observed CMB maps is an important issue while estimating a pure CMB signal. A component separation method, if these errors in the input foreground contaminated CMB maps are not properly taken into account, may lead to bias in the cleaned CMB map and estimated CMB angular power spectrum. But the inability to exactly determine the calibrati…
▽ More
Residual error in calibration coefficients corresponding to observed CMB maps is an important issue while estimating a pure CMB signal. A component separation method, if these errors in the input foreground contaminated CMB maps are not properly taken into account, may lead to bias in the cleaned CMB map and estimated CMB angular power spectrum. But the inability to exactly determine the calibration coefficients corresponding to each observed CMB map from any CMB experiment makes it very difficult to incorporate their exact and actual values in a component separation analysis. Hence the effect of any random and residual calibration error on the cleaned CMB map and its angular power spectrum of a component separation problem can only be understood by performing detailed Monte Carlo simulations. In this paper, we investigate the impact of using input foreground contaminated CMB maps with random calibration errors on posterior density of cleaned CMB map and theoretical CMB angular power spectrum over large angular scales of the sky following the Gibbs ILC method proposed by \cite{Sudevan:2018qyj}. By performing detailed Monte Carlo simulations of WMAP and Planck temperature anisotropy observations with calibration errors compatible with them we show that the best-fit map corresponding to posterior maximum is minimally biased in Gibbs ILC method by a CMB normalization bias and residual foreground bias. The bias in best-fit CMB angular power spectrum with respect to the case where no calibration error is present are $\sim 28 μK^2$ and $-4.7 μK^2$ respectively between $2 \le \ell \le 15$ and $16 \le \ell \le 32$. The calibration error induced error in best-fit power spectrum causes an overall $6\%$ increase of the net error when added in quadrature with the cosmic variance induced error.
△ Less
Submitted 9 January, 2020;
originally announced January 2020.
-
An Application of Global ILC Algorithm over Large Angular Scales to Estimate CMB Posterior Using Gibbs Sampling
Authors:
Vipin Sudevan,
Rajib Saha
Abstract:
In this work, we formalize a new technique to investigate joint posterior density of Cosmic Microwave Background (CMB) signal and its theoretical angular power spectrum given the observed data, using the global internal-linear-combination (ILC) method first proposed by Sudevan & Saha (2017). We implement the method on low resolution CMB maps observed by WMAP and Planck satellite missions, using Gi…
▽ More
In this work, we formalize a new technique to investigate joint posterior density of Cosmic Microwave Background (CMB) signal and its theoretical angular power spectrum given the observed data, using the global internal-linear-combination (ILC) method first proposed by Sudevan & Saha (2017). We implement the method on low resolution CMB maps observed by WMAP and Planck satellite missions, using Gibbs sampling, assuming that the detector noise is negligible on large angular scales of the sky. The main products of our analysis are best fit CMB cleaned map and its theoretical angular power spectrum along with their error estimates. We validate the methodology by performing Monte Carlo simulations that includes realistic foreground models and noise levels consistent with WMAP and Planck observations. Our method has an unique advantage that the posterior density is obtained without any need to explicitly model foreground components. Secondly, the power spectrum results with the error estimates can be directly used for cosmological parameter estimations.
△ Less
Submitted 20 October, 2018;
originally announced October 2018.
-
A Global ILC Approach in Pixel Space over Large Angular Scales of the Sky using CMB Covariance Matrix
Authors:
Vipin Sudevan,
Rajib Saha
Abstract:
We propose a new internal linear combination (ILC) method in the pixel space, applicable on large angular scales of the sky, to estimate a foreground minimized Cosmic Microwave Background (CMB) temperature anisotropy map by incorporating prior knowledge about the theoretical CMB covariance matrix. Usual ILC method in pixel space, on the contrary, does not use any information about the underlying C…
▽ More
We propose a new internal linear combination (ILC) method in the pixel space, applicable on large angular scales of the sky, to estimate a foreground minimized Cosmic Microwave Background (CMB) temperature anisotropy map by incorporating prior knowledge about the theoretical CMB covariance matrix. Usual ILC method in pixel space, on the contrary, does not use any information about the underlying CMB covariance matrix. The new approach complements the usual pixel space ILC technique specifically at low multipole region, using global information available from theoretical CMB covariance matrix as well as from the data. Since we apply our method over the large scale on the sky containing low multipoles we perform foreground minimization globally. We apply our methods on low resolution Planck and WMAP foreground contaminated CMB maps and validate the methodology by performing detailed Monte-Carlo simulations. Our cleaned CMB map and its power spectrum have significantly less error than those obtained following usual ILC technique at low resolution that does not use CMB covariance information. Another very important advantage of our method is that the cleaned power spectrum does not have any negative bias at the low multipoles because of effective suppression of CMB-foreground chance correlations on large angular scales of the sky. Our cleaned CMB map and its power spectrum match well with those estimated by other research groups.
△ Less
Submitted 28 December, 2017;
originally announced December 2017.
-
An Improved Diffuse Foreground Subtraction by ILC method: CMB Map and Angular Power Spectrum using Planck and WMAP Observations
Authors:
Vipin Sudevan,
Pavan K. Aluri,
Sarvesh Kumar Yadav,
Rajib Saha,
Tarun Souradeep
Abstract:
We report an improved technique for diffuse foreground minimization from Cosmic Microwave Background (CMB) maps using a new multi-phase iterative internal-linear-combination (ILC) approach in harmonic space. The new procedure consists of two phases. In phase 1, a diffuse foreground cleaned map is obtained by performing a usual ILC operation in the harmonic space in a single iteration over the desi…
▽ More
We report an improved technique for diffuse foreground minimization from Cosmic Microwave Background (CMB) maps using a new multi-phase iterative internal-linear-combination (ILC) approach in harmonic space. The new procedure consists of two phases. In phase 1, a diffuse foreground cleaned map is obtained by performing a usual ILC operation in the harmonic space in a single iteration over the desired portion of the sky. In phase 2, we obtain the final foreground cleaned map using an iterative ILC approach also in the harmonic space, however, now, during each iteration of foreground minimization, some of the regions of the sky that are not being cleaned in the current iteration, are replaced by the corresponding cleaned portions of the phase 1 cleaned map. The new ILC method nullifies a foreground leakage signal that is otherwise inevitably present in the old and usual harmonic space iterative ILC method. The new method is flexible to handle input frequency maps, irrespective of whether or not they initially have the same instrumental and pixel resolution, by bringing them to a common and maximum possible beam and pixel resolution at the beginning of the analysis. This dramatically reduces data redundancy and hence memory usage and computational cost. During the ILC weight calculation it avoids any need to deconvolve partial sky spherical harmonic coefficients by the beam and pixel window functions, which in strict mathematical sense, is not well-defined for azimuthally symmetric window functions. Using WMAP 9-year and Planck-2015 published frequency maps we obtain a pair of foreground cleaned CMB maps and CMB angular power spectrum. Our power spectrum match well with Planck-2015 results, with some difference. Finally, we show that the weights for ILC foreground minimization have an intrinsic characteristic that it tends to produce a statistically isotropic CMB map as well.
△ Less
Submitted 11 December, 2016;
originally announced December 2016.
