arXiv subjects

Supranta S. Boruah

Publications and source records attributed to Supranta S. Boruah.

At least 19 recordsLinked to original sources

arXiv2025

High-resolution weak lensing mass mapping from DES-Y3 data using diffusion-based prior

High-resolution mapping of cosmic mass distribution is essential for a variety of astrophysical applications including understanding cosmic structure formation, and galaxy formation and evolution. However dark matter is not directly observed and therefore we need advanced methods for solving inverse problems to reconstruct the underlying cosmic matter distribution. Here, we train a generative diffusion model and use it in the Diffusion Posterior Sampling (DPS) framework to reconstruct mass maps from Dark Energy Survey-Year 3 (DES-Y3) weak gravitational lensing data at high (1 arcminute) resolution. We show that the standard DPS results are biased, but they can be easily corrected by scaling the log-likelihood score during the diffusion process, yielding unbiased results with proper uncertainty quantification. The resulting mass maps reveal cosmic structures with enhanced detail, opening the door for improved astrophysical studies using the obtained mass maps.

astro-ph.CO↗

arXiv2025

Inference of Hubble constant using standard sirens and reconstructed matter density field

We summarise a new approach for measuring the Hubble constant using standard sirens and the reconstructed matter density field obtained from observed galaxy surveys. Specifically, we describe and test this method using the Bayesian forward-modelled software BORG. This software evolves the initial density field to the present, sampling plausible density fields from the posterior, accounting for peculiar velocities, and automatically incorporating higher-order correlation functions. The advantage of adding additional information from correlations is expected to make this method more effective in low-signal-to-noise regimes, such as those with modest galaxy number density or incomplete surveys. Our results show that developing a cross-correlation framework between gravitational waves and galaxy surveys, based on the forward-modelled reconstructed density field, to measure H0 is promising. This review is based on the research conducted as part of Boruah's Ph.D. thesis.

astro-ph.CO↗

arXiv2025

Diffusion-based mass map reconstruction from weak lensing data

Diffusion models have been used in cosmological applications as a generative model for fast simulations and to reconstruct underlying cosmological fields or astrophysical images from noisy data. These two tasks are often treated as separate: diffusion models trained for one purpose do not generalize to perform the other task. In this paper, we develop a single diffusion model that can be used for both tasks. By using the Diffusion Posterior Sampling (DPS) approach, we use a diffusion model trained to simulate weak lensing maps for the inverse problem of reconstructing mass maps from noisy weak lensing data. We find that the standard DPS method leads to biased inference but we correct this bias by down weighting the likelihood term at early sampling time steps of the diffusion. Our method give us a way to reconstruct accurate high-resolution (sub-arcminute) mass maps that have the correct power spectrum and a range of non-Gaussian summary statistics. We discuss several applications enabled by the computational efficiency and accuracy of our model. These include generation of simulation quality mass maps, aiding covariance estimation for higher order statistics, and for finding filaments, voids and clusters from noisy lensing shear data.

astro-ph.CO↗

arXiv2024

Fast Generation of Weak Lensing Maps with Analytical Point Transformation Functions

Nonlinear cosmological fields like galaxy density and lensing convergence can be approximately related to Gaussian fields via analytic point transforms. The lognormal transform (LN) has been widely used and is a simple example of a function that relates nonlinear fields to Gaussian fields. We consider more accurate General Point-Transformed Gaussian (GPTG) functions for such a mapping and apply them to convergence maps. We show that we can create maps that preserve the LN's ability to exactly match any desired power spectrum but go beyond LN by significantly improving the accuracy of the probability distribution function (PDF). With the aid of symbolic regression, we find a remarkably accurate GPTG function for convergence maps: its higher-order moments, scattering wavelet transform, Minkowski functionals, and peak counts match those of N-body simulations to the statistical uncertainty expected from tomographic lensing maps of the Rubin LSST 10 years survey. Our five-parameter function performs 2 to 5$\times$ better than the lognormal. We restrict our study to scales above about 7 arcmin; baryonic feedback alters the mass distribution on smaller scales. We demonstrate that the GPTG can robustly emulate variations in cosmological parameters due to the simplicity of the analytic transform. This opens up several possible applications, such as field-level inference, rapid covariance estimation, and other uses based on the generation of arbitrarily many maps with laptop-level computation capability.

astro-ph.CO↗

arXiv2024

GANSky -- fast curved sky weak lensing simulations using Generative Adversarial Networks

Extracting non-Gaussian information from the next generation weak lensing surveys will require fast and accurate full-sky simulations. This is difficult to achieve in practice with existing simulation methods: ray-traced $N$-body simulations are computationally expensive, and approximate simulation methods (such as lognormal mocks) are not accurate enough. Here, we present GANSky, an interpretable machine learning method that uses Generative Adversarial Networks (GANs) to produce fast and accurate full-sky tomographic weak lensing maps. The input to our GAN are lognormal maps that approximately describe the late-time convergence field of the Universe. Starting from these lognormal maps, we use GANs to learn how to locally redistribute mass to achieve simulation-quality maps. This can be achieved using remarkably small networks ($\approx 10^3$ parameters). We validate the GAN maps by computing a number of summary statistics in both simulated and GANSky maps. We show that GANSky maps correctly reproduce both the mean and $χ^2$ distribution for several statistics, specifically: the 2-pt function, 1-pt PDF, peak and void counts, and the equilateral, folded and squeezed bispectra. These successes makes GANSky an attractive tool to compute the covariances of these statistics. In addition to being useful for rapidly generating large ensembles of artificial data sets, our method can be used to extract non-Gaussian information from weak lensing data with field-level or simulation-based inference.

astro-ph.CO↗

arXiv2024

Attention-Based Neural Network Emulators for Multi-Probe Data Vectors Part II: Assessing Tension Metrics

The next generation of cosmological surveys is expected to generate unprecedented high-quality data, consequently increasing the already substantial computational costs of Bayesian statistical methods. This will pose a significant challenge to analyzing theoretical models of cosmology. Additionally, new mitigation techniques of baryonic effects, intrinsic alignment, and other systematic effects will inevitably introduce more parameters, slowing down the convergence of Bayesian analyses. In this scenario, machine-learning-based accelerators are a promising solution, capable of reducing the computational costs and execution time of such tools by order of thousands. Yet, they have not been able to provide accurate predictions over the wide prior ranges in parameter space adopted by Stage III/IV collaborations in studies employing real-space two-point correlation functions. This paper offers a leap in this direction by carefully investigating the modern transformer-based neural network (NN) architectures in realistic simulated Rubin Observatory year one cosmic shear $Λ$CDM inferences. Building on the framework introduced in Part I, we generalize the transformer block and incorporate additional layer types to develop a more versatile architecture. We present a scalable method to efficiently generate an extensive training dataset that significantly exceeds the scope of prior volumes considered in Part I, while still meeting strict accuracy standards. Through our meticulous architecture comparison and comprehensive hyperparameter optimization, we establish that the attention-based architecture performs an order of magnitude better in accuracy than widely adopted NN designs. Finally, we test and apply our emulators to calibrate tension metrics.

astro-ph.CO↗

arXiv2024

Machine Learning LSST 3x2pt analyses -- forecasting the impact of systematics on cosmological constraints using neural networks

Validating modeling choices through simulated analyses and quantifying the impact of different systematic effects will form a major computational bottleneck in the preparation for 3$\times$2 analysis with Stage-IV surveys such as Vera Rubin Observatory's Legacy Survey of Space and Time (LSST). We can significantly reduce the computational requirements by using machine learning based emulators, which allow us to run fast inference while maintaining the full realism of the data analysis pipeline. In this paper, we use such an emulator to run simulated 3$\times$2 (cosmic shear, galaxy-galaxy lensing, and galaxy clustering) analyses for mock LSST-Y1/Y3/Y6/Y10 surveys and study the impact of various systematic effects (galaxy bias, intrinsic alignment, baryonic physics, shear calibration and photo-$z$ uncertainties). Closely following the DESC Science Requirement Document (with several updates) our main findings are: {\it a)} The largest contribution to the `systematic error budget' of LSST 3$\times$2 analysis comes from galaxy bias uncertainties, while the contribution of baryonic and shear calibration uncertainties are significantly less important. {\it b)} Tighter constraints on intrinsic alignment and photo-$z$ parameters can improve cosmological constraints noticeably, which illustrates synergies of LSST and spectroscopic surveys. {\it c)} The scale cuts adopted in the DESC SRD may be too conservative and pushing to smaller scales can increase cosmological information significantly. {\it d)} We investigate the impact of photo-$z$ outliers on 3$\times$2 pt analysis and find that we need to determine the outlier fraction to within $5-10\%$ accuracy to ensure robust cosmological analysis. We caution that these findings depend on analysis choices (parameterizations, priors, scale cuts) and can change for different settings.

astro-ph.CO↗

arXiv2024

Bayesian mass mapping with weak lensing data using KARMMA -- validation with simulations and application to Dark Energy Survey Year 3 data

We update the field-level inference code KARMMA to enable tomographic forward-modelling of shear maps. Our code assumes a lognormal prior on the convergence field, and properly accounts for the cross-covariance in the lensing signal across tomographic source bins. We use mock weak lensing data from N-body simulations to validate our mass-mapping forward model by comparing our posterior maps to the input convergence fields. We find that KARMMA produces more accurate reconstructions than traditional mass-mapping algorithms. More-over, the KARMMA posteriors reproduce all statistical properties of the input density field we tested -- one- and two-point functions, and the peak and void number counts -- with $\lesssim~10\%$ accuracy. Our posteriors exhibit a small bias that increases with decreasing source redshift, but these biases are small compared to the statistical uncertainties of current (DES) cosmic shear surveys. Finally, we apply KARMMA to Dark Energy Survey Year 3 (DES-Y3) weak lensing data, and verify that the two point shear correlation function $ξ_+$ is well fit by the correlation function of the reconstructed convergence field. This is a non-trivial test that traditional mass mapping algorithms fail. The code is publicly available at https://github.com/Supranta/KaRMMa.git. KARMMA DES-Y3 mass maps are publicly available at https://zenodo.org/records/10672062.

astro-ph.CO↗

arXiv2024

Attention-based Neural Network Emulators for Multi-Probe Data Vectors Part I: Forecasting the Growth-Geometry split

We present a new class of machine-learning emulators that accurately model the cosmic shear, galaxy-galaxy lensing, and galaxy clustering real space correlation functions in the context of Rubin Observatory year one simulated data. To illustrate its capabilities in forecasting models beyond the standard $Λ$CDM, we forecast how well LSST Year 1 data will be able to probe the consistency between geometry $Ω^{\rm geo}_\mathrm{m}$ and growth $Ω^{\rm growth}_\mathrm{m}$ dark matter densities in the so-called split $Λ$CDM parameterization. When trained with a few million samples, our emulator shows uniform accuracy across a wide range in an 18-dimensional parameter space. We provide a detailed comparison of three neural network designs, illustrating the importance of adopting state-of-the-art Transformer blocks. Our study also details their performance when computing Bayesian evidence for cosmic shear on three fiducial cosmologies. The transformers-based emulator is always accurate within PolyChord's precision. As an application, we use our emulator to study the degeneracies between dark energy models and growth geometry split parameterizations. We find that the growth-geometry split remains to be a meaningful test of the smooth dark energy assumption.

astro-ph.CO↗

arXiv2023

Map-based cosmology inference with weak lensing -- information content and its dependence on the parameter space

Field-level inference is emerging as a promising technique for optimally extracting information from cosmological datasets. Indeed, previous analyses have shown field-based inference produces tighter parameter constraints than power spectrum analyses. However, estimates of the detailed quantitative gain in constraining power differ. Here, we demonstrate the gain in constraining power depends on the parameter space being constrained. As a specific example, we find that field-based analysis of an LSST Y1-like mock data set only marginally improves constraints relative to a 2-point function analysis in $Λ$CDM, yet it more than doubles the constraining power of the data in the context of $w$CDM models. This effect reconciles some, but not all, of the discrepant results found in the literature. Our results demonstrate the importance of using a full systematics model when quantifying the information gain for realistic field-level analyses of future data sets.

astro-ph.CO↗

arXiv2022

KaRMMa 2.0 -- Kappa Reconstruction for Mass Mapping

We present KaRMMa 2.0, an updated version of the mass map reconstruction code introduced in Fiedorowicz et al. (2022). KaRMMa is a full-sky Bayesian algorithm for reconstructing weak lensing mass maps from shear data. It forward-models the convergence field as a realization of a lognormal field. The corresponding shear map is calculated using the standard Kaiser-Squires transformation, and compared to observations at the field level. The posterior distribution of maps given the shear data is sampled using Hamiltonian Monte Carlo chains. Our work improves on the original algorithm by making it numerically efficient, enabling full-sky reconstructions at $\approx$ 7 arcmin resolution with modest computational resources. These gains are made with no loss in accuracy or precision relative to KaRMMa 1.0. We compare the KaRMMa 2.0 posteriors against simulations across a variety of summary statistics (one-point function, two-point functions, and peak/void counts) to demonstrate our updated algorithm provides an accurate reconstruction of the convergence field at mildly non-linear scales. Unsurprisingly, the lognormal model fails as we approach non-linear scales ($\ell \gtrsim 200$), which in turn biases the map posteriors. These biases are at the 2% level in the recovered power spectrum, and at the 5% to 15% level for other statistics, depending on the resolution.

astro-ph.CO↗

arXiv2022

Accelerating cosmological inference with Gaussian processes and neural networks -- an application to LSST Y1 weak lensing and galaxy clustering

Studying the impact of systematic effects, optimizing survey strategies, assessing tensions between different probes and exploring synergies of different data sets require a large number of simulated likelihood analyses, each of which cost thousands of CPU hours. In this paper, we present a method to accelerate cosmological inference using emulators based on Gaussian process regression and neural networks. We iteratively acquire training samples in regions of high posterior probability which enables accurate emulation of data vectors even in high dimensional parameter spaces. We showcase the performance of our emulator with a simulated 3x2 point analysis of LSST-Y1 with realistic theoretical and systematics modelling. We show that our emulator leads to high-fidelity posterior contours, with an order of magnitude speed-up. Most importantly, the trained emulator can be re-used for extremely fast impact and optimization studies. We demonstrate this feature by studying baryonic physics effects in LSST-Y1 3x2 point analyses where each one of our MCMC runs takes approximately 5 minutes. This technique enables future cosmological analyses to map out the science return as a function of analysis choices and survey strategy.

astro-ph.CO↗

arXiv2022

KaRMMa -- Kappa Reconstruction for Mass Mapping

We present KaRMMa, a novel method for performing mass map reconstruction from weak-lensing surveys. We employ a fully Bayesian approach with a physically motivated lognormal prior to sample from the posterior distribution of convergence maps. We test KaRMMa on a suite of dark matter N-body simulations with simulated DES Y1-like shear observations. We show that KaRMMa outperforms the basic Kaiser-Squires mass map reconstruction in two key ways: 1) our best map point estimate has lower residuals compared to Kaiser-Squires; and 2) unlike the Kaiser-Squires reconstruction, the posterior distribution of KaRMMa maps are nearly unbiased in all summary statistics we considered, namely: one-point and two-point functions, and peak/void counts. In particular, KaRMMa successfully captures the non-Gaussian nature of the distribution of $κ$ values in the simulated maps. We further demonstrate that the KaRMMa posteriors correctly characterize the uncertainty in all summary statistics we considered.

astro-ph.CO↗

arXiv2021

Reconstructing dark matter distribution with peculiar velocities: Bayesian forward modelling with corrections for inhomogeneous Malmquist bias

We present a forward-modelled velocity field reconstruction algorithm that performs the reconstruction of the mass density field using only peculiar velocity data. Our method consistently accounts for the inhomogeneous Malmquist bias using analytic integration along the line-of-sight. By testing our method on a simulation, we show that our method gives an unbiased reconstruction of the velocity field. We show that not accounting for the inhomogeneous Malmquist bias can lead to significant biases in the forward-modelled reconstructions. We applied our method to a peculiar velocity data set consisting of the SFI++ and 2MTF Tully-Fisher catalogues and the A2 supernovae compilation, thus obtaining a novel velocity reconstruction in the local Universe. Our velocity reconstructions have a cosmological power spectrum consistent with the theoretical expectation. Furthermore, we obtain a full description of the uncertainties on reconstruction through samples of the posterior distribution. We validate our velocity reconstruction of the local Universe by comparing it to an independent reconstruction using the 2M++ galaxy catalogue, obtaining good agreement between the two reconstructions. Using Bayesian model comparison, we find that our velocity model performs better than the adaptive kernel smoothed velocity with the same peculiar velocity data. However, our velocity model does not perform as well as the velocity reconstruction from the 2M++ galaxy catalogue, due to the sparse and noisy nature of the peculiar velocity tracer samples. The method presented here provides a way to include peculiar velocity data in initial condition reconstruction frameworks.

astro-ph.CO↗

arXiv2021

Peculiar-velocity cosmology with Types Ia and II supernovae

We present the Democratic Samples of Supernovae (DSS), a compilation of 775 low-redshift Type Ia and II supernovae (SNe Ia & II), of which 137 SN Ia distances are derived via the newly developed snapshot distance method. Using the objects in the DSS as tracers of the peculiar-velocity field, we compare against the corresponding reconstruction from the 2M++ galaxy redshift survey. Our analysis -- which takes special care to properly weight each DSS subcatalogue and cross-calibrate the relative distance scales between them -- results in a measurement of the cosmological parameter combination $fσ_8 = 0.390_{-0.022}^{+0.022}$ as well as an external bulk flow velocity of $195_{-23}^{+22}$ km s$^{-1}$ in the direction $(\ell, b) = (292_{-7}^{+7}, -6_{-4}^{+5})$ deg, which originates from beyond the 2M++ reconstruction. Similarly, we find a bulk flow of $245_{-31}^{+32}$ km s$^{-1}$ toward $(\ell, b) = (294_{-7}^{+7}, 3_{-5}^{+6})$ deg on a scale of $\sim 30 h^{-1}$ Mpc if we ignore the reconstructed peculiar-velocity field altogether. Our constraint on $fσ_8$ -- the tightest derived from SNe to date (considering only statistical error bars), and the only one to utilise SNe II -- is broadly consistent with other results from the literature. We intend for our data accumulation and treatment techniques to become the prototype for future studies that will exploit the unprecedented data volume from upcoming wide-field surveys.

astro-ph.CO↗

arXiv2020

Cosmic flows in the nearby Universe: new peculiar velocities from SNe and cosmological constraints

The peculiar velocity field offers a unique way to probe dark matter density field on large scales at low redshifts. In this work, we have compiled a new sample of 465 peculiar velocities from low redshift $(z < 0.067)$ Type Ia supernovae. We compare the reconstructed velocity field derived from the 2M++ galaxy redshift compilation to the supernovae, the SFI++ and the 2MTF Tully-Fisher distance catalogues. We used a forward method to jointly infer the distances and the velocities of distance indicators by comparing the observations to the reconstruction. Comparison of the reconstructed peculiar velocity fields to observations allows us to infer the cosmological parameter combination $fσ_8$, and the bulk flow velocity arising from outside the survey volume. The residual bulk flow arising from outside the 2M++ volume is inferred to be $171^{+11}_{-11}$ km s$^{-1}$ in the direction $l=301^{\circ} \pm 4^{\circ}$ and $b=0^{\circ} \pm 3^{\circ}$. We obtain $fσ_{8} = 0.400 \pm 0.017$, equivalent to $S_8 \approx σ_8(Ω_m/0.3)^{0.55}=0.776\pm0.033$, which corresponds to an approximately $ 4\%\,$ statistical uncertainty on the value of $fσ_8$. Our inferred value is consistent with other low redshift results in the literature.

astro-ph.CO↗

arXiv2020

Peculiar velocities in the local Universe: comparison of different models and the implications for $H_0$ and dark matter

When measuring the value of the Hubble parameter, $H_0$, it is necessary to know the recession velocity free of the effects of peculiar velocities. In this work, we study different models of peculiar velocity in the local Universe. In particular, we compare models based on density reconstruction from galaxy redshift surveys and kernel smoothing of peculiar velocity data. The velocity field from the density reconstruction is obtained using the 2M++ galaxy redshift compilation, which is compared to two adaptive kernel-smoothed velocity fields: the first obtained from the 6dF Fundamental Plane sample and the other using a Tully-Fisher catalogue obtained by combining SFI++ and 2MTF. We highlight that smoothed velocity fields should be rescaled to obtain unbiased velocity estimates. Comparing the predictions of these models to the observations from a few test sets of peculiar velocity data, obtained from the Second Amendment Supernovae catalogue and the Tully-Fisher catalogues, we find that 2M++ reconstruction provides a better model of the peculiar velocity in the local Universe than the kernel-smoothed peculiar velocity models. We study the impact of peculiar velocities on the measurement of $H_0$ from gravitational waves and megamasers. In doing so, we introduce a probabilistic framework to marginalize over the peculiar velocity corrections along the line-of-sight. For the megamasers, we find $H_0 = 69^{+2.9}_{-2.8}$ km s^{-1} Mpc^{-1} using the 2M++ velocity field. We also study the peculiar velocity of the the galaxy NGC1052-DF2, concluding that a short $\sim$ 13 Mpc distance is not a likely explanation of the anomalously low dark matter fraction of that galaxy.

astro-ph.CO↗

arXiv2018

Cuscuton Bounce

In general relativity producing a regular bounce entails violation of Null Energy Condition for a dynamical source in the model. That generically indicates existence of ghosts or other instabilities. However, in cuscuton modification of gravity, the correspondence between a background bounce and violation of Null Energy Condition for dynamical sources is broken. Cuscuton is an Infra Red modification and it can make the background go through a regular bounce. At the same time, since it does not contain any dynamical degrees of freedom, it does not lead to ghosts or other instabilities. Here, we present a toy scenario of a regular bouncing cosmology and prove this claim. Our model is presented as a proof of concept at this point and does not aim to explain observations in late time cosmology.

gr-qc↗