arXiv subjects

Ayan Mitra

Publications and source records attributed to Ayan Mitra.

At least 19 recordsLinked to original sources

arXiv2026

Fisher Forecasting for the DESC with $\texttt{Augur}$

The Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST) has begun its ten-year survey of the entire visible southern hemisphere. To ensure robust cosmological measurements, computationally inexpensive investigations of modeling choices must be made to gauge the performance of proposed cosmological analyses. In this paper, we introduce the $\texttt{Augur}$ tool of the Dark Energy Science Collaboration (DESC), which provides Fisher forecasts for cosmological inference for the LSST using software frameworks designed for DESC science. We test the pipeline by comparing it to forecasts produced by external code and direct sampling of the posterior via nested sampling methods, finding good agreement between all methods. We additionally investigate a range of modeling and hyperparameter choices for a 3$\times$2pt investigation in harmonic space, providing users with diagnostics to obtain reliable forecasts. $\texttt{Augur}$ will be continually updated to be compatible with the other tools in the DESC software ecosystem as additional probes and functionality become available.

astro-ph.CO↗

arXiv2026

Machine Learning Closure Audits for LSST Photometric Supernova Cosmology

Modern and next generation supernova cosmology analyses rely on end to end simulations to train photometric classifiers, characterise selection effects, validate light curve models, and calibrate distance bias corrections. Standard closure tests based on global Hubble diagram summaries can miss multivariate structure that remains in post correction residuals. We introduce a supervised machine learning closure audit that tests whether measured observables can predict the bias corrected Hubble residuals $\Delta\mu$. We apply the audit to LSST Type Ia supernova simulations from two independent analyses: the M23 mock data sets of Mitra et al. (2023), with spectroscopic redshift and photometric redshift samples, and the predominantly photometric LSST like M25 simulation of Mitra et al. (2025). Standard one dimensional Redshift binned diagnostics explain <1% of the residual variance ($R^2 < 0.01$), suggesting apparent closure. In contrast, out of fold LightGBM models recover up to 98.2% of the variance in the simulated residuals ($R^2 = 0.982$), revealing structured residual predictability. Applying the same audit directly to the real Dark Energy Survey 5 Year (DES 5YR) spectroscopic sample yields a held out $R^2 = 0.725$. SHAP feature attribution rankings are highly consistent between independently trained M25 and DES models (Spearman $\rho = 0.802$), with apparent magnitudes, signal to noise ratios, and redshift dominating the shared predictive hierarchy. The resulting scorecard provides a diagnostic protocol for comparing mock ensembles and real observations, identifying residual non closure before cosmological parameters are unblinded.

astro-ph.CO↗

arXiv2026

Catching the Cosmic Sign Flip: Background and Growth Tests of Smooth Sign Switching Lambda_s CDM

Recent baryon acoustic oscillation (BAO) measurements from the Dark Energy Spectroscopic Instrument (DESI), in combination with CMB and supernova data, show a mild preference for dynamical dark energy over flat $\Lambda$CDM. They can also drive the best-fit effective neutrino mass to unphysical negative values, motivating models that mimic this effect through late-time expansion physics. The sign-switching cosmological constant ($\Lambda_s$CDM) model addresses this by introducing an AdS-like negative vacuum energy density at high redshifts ($z > z^\dagger$) that transitions to a dS-like positive density at low redshifts. While original fits assumed a discontinuous step function transition, we generalise the dynamics using a smooth hyperbolic tangent parametrisation, $\Lambda(z) \propto \tanh[\Delta(z^\dagger - z)]$. We constrain the transition redshift $z^\dagger$ and smoothness parameter $\Delta$ using background distance data from DES-SN5YR (with Dovekie recalibration) and DESI 2024 BAO. We further incorporate a compact redshift space distortion (RSD) $f\sigma_8$ compilation. The joint data do not prefer the smooth sign-switching model over flat $\Lambda$CDM, yielding nearly identical $\chi^2_{\text{min}}$ values, while information criteria penalise the additional parameters. The transition smoothness remains unconstrained and dominated by the prior volume ($\Delta = 24.75^{+17.42}_{-17.48}$), indicating that current data cannot distinguish between a sharp transition and a smooth dynamical crossover. This work does not directly constrain the physical sum of neutrino masses $\sum m_\nu$; rather, it shows that current late-time background and growth data alone cannot resolve the transition details. We discuss the implications of this null detection and present projections for next-generation surveys.

astro-ph.CO↗

arXiv2026

The Generalization Gap in Machine Learning EoS Inference from Core-Collapse Supernova Gravitational Waves

Core-collapse supernova gravitational waves may carry information about the dense matter equation of state (EoS), which describes the relation between pressure, density, temperature, and composition. This work tests a crucial question for physical inference: can a machine learning model trained on a finite simulation catalogue predict EoS parameters for an EoS family that was absent during training? Under standard random cross-validation, a LightGBM regressor appears highly successful, yielding $R^2=(0.70,0.67,0.60)$ for the nuclear incompressibility, symmetry energy, and slope parameter $(K_0,J,L)$. However, under Leave-One-EoS-Out (LOEO) validation, where all waveforms from a single EoS are withheld, the model fails, yielding mean absolute errors of $(44.57,3.19,30.54)$ MeV and negative pooled $R^2$ scores, performing worse than a baseline mean predictor. This generalisation gap persists across linear models, random forests, neural networks, and gradient-boosted trees. Restricting inputs to physical features (bounce amplitude, bounce width, peak frequency) reduces template leakage, the memorisation of related templates shared across random splits, but does not restore reliable EoS extrapolation. In contrast, a progenitor mass case study shows that classification generalises to unseen rotation speeds, while continuous mass regression compresses predictions towards the catalogue interior. These results demonstrate that while machine learning successfully interpolates within current waveform catalogues, this does not imply robust physical inference for unseen EoS models. Future pipelines should adopt leave-family-out validation, wider simulation coverage, and physics-aware inference frameworks.

astro-ph.HE↗

arXiv2026

FlowSN: Neural Simulation-Based Inference under Realistic Selection Effects applied to Supernova Cosmology

We present FlowSN, a statistical framework using simulation-based inference (SBI) with normalising flows to account for selection effects in observational astronomy. Failure to account for selection effects can lead to biased inference on global parameters. An example is Malmquist bias, where detection limits result in a sample skewed towards brighter objects. In Type Ia supernova (SN Ia) cosmology, these selection effects can systematically shift the inferred posterior distributions of cosmological parameters, necessitating the development of robust statistical frameworks to account for the biases. SBI enables us to implicitly learn probability distributions that are analytically intractable to calculate. In this work, we introduce a novel approach that employs a normalising flow to learn the non-analytic selected SN likelihood for a given survey from forward simulations, independent of the assumed cosmological model. The resulting likelihood approximation is incorporated into a hierarchical Bayesian framework and posterior sampling is performed using Hamiltonian Monte Carlo to obtain constraints on cosmological parameters conditioned on the observed data. The modular learnt likelihood approximation can be reused without retraining to evaluate different cosmological models, providing a key advantage over other SBI approaches. We demonstrate the performance of this methodology by training and testing the SBI technique using realistic LSST-like SNANA simulations for the first time. Our FlowSN approach yields accurate posterior estimates on cosmological parameters, including the dark energy equation of state $w_0$, that are an order of magnitude less biased than those obtained with conventional techniques and also exhibit improved frequentist calibration.

astro-ph.CO↗

arXiv2026

Probing Physics Beyond the Standard Model through Combined Analyses of Next-Generation Type Ia Supernova, CMB, and BAO Surveys

Observations of Type Ia supernovae (\sne), which probe the late Universe, together with baryon acoustic oscillations (BAO) and the cosmic microwave background (CMB), which probe the intermediate and early epochs, provide complementary constraints on the expansion history of the Universe. In this work, we forecast constraints on dark energy and other extensions to the standard cosmological model by combining the SNIa sample expected from the Vera C. Rubin Observatory's Legacy Survey of Space and Time (LSST), data from current and forthcoming CMB surveys, and BAO measurements from the Dark Energy Spectroscopic Instrument (DESI). For the CMB, we use temperature, polarization, and lensing power spectra ($TT/EE/TE/\phi\phi$) from South Pole Telescope, the planned Advanced Simons Observatory, and a CMB-S4-like experiment. We derive constraints on $\Lambda {\rm CDM}$ and its extensions involving the dark energy equation of state parameters $(w_{0}, w_{a})$ and the sum of neutrino masses $\sum m_{\nu}$, using a Markov Chain Monte Carlo (MCMC) sampling framework. We find that the LSST Year-3 SNIa sample can improve upon the DES Year-5 dark energy constraints by a factor of $\times2-\times2.5$, with the gains driven primarily by the significantly higher SNIa density in the LSST sample. Similarly, DESI-DR3 shows up to a $\times1.8$ improvement on dark energy parameters over DR2, driven largely by the substantial increase in low-redshift sample. Combining CMB with LSST-Y3-SNIa and DESI-DR3-BAO yields $\sigma(w_{0}) = 0.028$ and $\sigma(w_{a}) = 0.11$ for $w_{0} w_{a} {\rm CDM}$ cosmology with the results being largely independent of the CMB dataset. The constraints weaken by 10%-30% when freeing $\sum m_{\nu}$ and spatial curvature. Moreover, the joint analysis of the three datasets can enable a $2-3\sigma$ detection of $\sum m_{\nu}$.

astro-ph.CO↗

arXiv2025

A Fully Photometric Approach to Type Ia Supernova Cosmology in the LSST Era: Host Galaxy Redshifts and Supernova Classification

The upcoming Vera C. Rubin Observatory's Legacy Survey of Space and Time (LSST) is expected to discover nearly a million Type Ia supernovae (SNeIa), offering an unprecedented opportunity to constrain dark energy. The vast majority of these events will lack spectroscopic classification and redshifts, necessitating a fully photometric approach to maximize cosmology constraining power. We present detailed simulations based on the Extended LSST Astronomical Time Series Classification Challenge (ELAsTiCC), and a cosmological analysis using photometrically classified SNeIa with host galaxy photometric redshifts. This dataset features realistic multi-band light curves, non-SNIa contamination, host mis-associations, and transient-host correlations across the high-redshift Deep Drilling Fields (DDF) (~ 50 deg^2). We also include a spectroscopically confirmed low-redshift sample based on the Wide Fast Deep (WFD) fields. We employ a joint SN+host photometric redshift fit, a neural network based photometric classifier (SCONE), and BEAMS with Bias Corrections (BBC) methodology to construct a bias-corrected Hubble diagram. We produce statistical + systematic covariance matrices, and perform cosmology fitting with a prior using Cosmic Microwave Background constraints. We fit and present results for the wCDM dark energy model, and the more general Chevallier-Polarski-Linder (CPL) w0wa model. With a simulated sample of ~6000 events, we achieve a Figure of Merit (FoM) value of about 150, which is significantly larger than the DESVYR FoM of 54. Averaging analysis results over 25 independent samples, we find small but significant biases indicating a need for further analysis testing and development.

astro-ph.CO↗

arXiv2025

Lens Model Accuracy in the Expected LSST Lensed AGN Sample

Strong gravitational lensing of active galactic nuclei (AGN) enables measurements of cosmological parameters through time-delay cosmography (TDC). With data from the upcoming LSST survey, we anticipate using a sample of O(1000) lensed AGN for TDC. To prepare for this dataset and enable this measurement, we construct and analyze a realistic mock sample of 1300 systems drawn from the OM10 (Oguri & Marshall 2010) catalog of simulated lenses with AGN sources at $z<3.1$ in order to test a key aspect of the analysis pipeline, that of the lens modeling. We realize the lenses as power law elliptical mass distributions and simulate 5-year LSST i-band coadd images. From every image, we infer the lens mass model parameters using neural posterior estimation (NPE). Focusing on the key model parameters, $θ_E$ (the Einstein Radius) and $γ_{lens}$ (the projected mass density profile slope), with consistent mass-light ellipticity correlations in test and training data, we recover $θ_E$ with less than 1% bias per lens, 6.5% precision per lens and $γ_{lens}$ with less than 3% bias per lens, 8% precision per lens. We find that lens light subtraction prior to modeling is only useful when applied to data sampled from the training prior. If emulated deconvolution is applied to the data prior to modeling, precision improves across all parameters by a factor of 2. Finally, we combine the inferred lens mass models using Bayesian Hierarchical Inference to recover the global properties of the lens sample with less than 1% bias.

astro-ph.CO↗

arXiv2025

Litmus tests of the flat $Λ$CDM model and model-independent measurement of $H_0r_\mathrm{d}$ with LSST and DESI

In this analysis we apply a model-independent framework to test the flat $Λ$CDM cosmology using simulated SNIa data from the upcoming Legacy Survey of Space and Time (LSST) and combined with simulated Dark Energy Spectroscopic Instrument (DESI) five-years Baryon Acoustic Oscillations (BAO) data. We adopt an iterative smoothing technique to reconstruct the expansion history from SNIa data, which, when combined with BAO measurements, facilitates a comprehensive test of the Universe's curvature and the nature of dark energy. The analysis is conducted under four different mock true cosmologies: Two curvatures ($Ω_{k,0}=0$ and 0.1) and two models of dark energy: a cosmological constant $Λ$ and the phenomenologically emergent dark energy. We forecast that our reconstruction technique can constrain cosmological parameters, such as the curvature ($Ω_{k,0}$) and $c/(H_0 r_\mathrm{d})$, with spread due to the SNIa uncertainties up to $\pm 4\%$ and $\pm 0.1$ respectively, without assuming any form of dark energy.

astro-ph.CO↗

arXiv2025

Evaluating Machine Learning Models for Supernova Gravitational Wave Signal Classification

We investigate the potential of using gravitational wave (GW) signals from rotating core-collapse supernovae to probe the equation of state (EOS) of nuclear matter. By generating GW signals from simulations with various EOSs, we train machine learning models to classify them and evaluate their performance. Our study builds on previous work by examining how different machine learning models, parameters, and data preprocessing techniques impact classification accuracy. We test convolutional and recurrent neural networks, as well as six classical algorithms: random forest, support vector machines, naïve Bayes, logistic regression, $k$-nearest neighbors, and eXtreme gradient boosting. All models, except naïve Bayes, achieve over 90 per cent accuracy on our dataset. Additionally, we assess the impact of approximating the GW signal using the general relativistic effective potential (GREP) on EOS classification. We find that models trained on GREP data exhibit low classification accuracy. However, normalizing time by the peak signal frequency, which partially compensates for the absence of the time dilation effect in GREP, leads to a notable improvement in accuracy. Despite this, the accuracy does not exceed 70 per cent, suggesting that GREP lacks the precision necessary for EOS classification. Finally, our study has several limitations, including the omission of detector noise and the focus on a single progenitor mass model, which will be addressed in future works.

astro-ph.HE↗

arXiv2024

Dark energy reconstruction analysis with artificial neural networks: Application on simulated Supernova Ia data from Rubin Observatory

In this paper, we present an analysis of Supernova Ia (SNIa) distance moduli $μ(z)$ and dark energy using an Artificial Neural Network (ANN) reconstruction based on LSST simulated three-year SNIa data. The ANNs employed in this study utilize genetic algorithms for hyperparameter tuning and Monte Carlo Dropout for predictions. Our ANN reconstruction architecture is capable of modeling both the distance moduli and their associated statistical errors given redshift values. We compare the performance of the ANN-based reconstruction with two theoretical dark energy models: $Λ$CDM and Chevallier-Linder-Polarski (CPL). Bayesian analysis is conducted for these theoretical models using the LSST simulations and compared with observations from Pantheon and Pantheon+ SNIa real data. We demonstrate that our model-independent ANN reconstruction is consistent with both theoretical models. Performance metrics and statistical tests reveal that the ANN produces distance modulus estimates that align well with the LSST dataset and exhibit only minor discrepancies with $Λ$CDM and CPL.

astro-ph.CO↗

arXiv2024

Exploring Alternative Cosmologies with the LSST: Simulated Forecasts and Current Observational Constraints

In recent years, the Lambda Cold Dark Matter (LCDM) model, which has been pivotal in cosmological studies, has faced significant challenges due to emerging observational and theoretical inconsistencies. This paper explores alternative cosmological models to address these discrepancies, using simulated three years photometric Supernovae Ia data from the Legacy Survey of Space and Time (LSST), supplemented with additional Pantheon+, Union, and the recently released Dark Energy Survey 5 Years (DESY5) supernova compilations and Baryon Acoustic Oscillation (BAO) measurements. We assess the constraining power of these datasets on various dynamic dark energy models, including CPL, BA, JBP, SCPL, and GCG. Our analysis demonstrates that the LSST with its high precision data, can provide tighter constraints on dark energy parameters compared to other datasets. Additionally, the inclusion of BAO measurements significantly improves parameter constraints across all models. Except for Pantheon+, we find that across all the cosmological datasets, and the dark energy models considered in this work, there is a consistent deviation from the LCDM model that exceeds a 2-sigma significance level. Our findings underscore the necessity of exploring dynamic dark energy models, which offer more consistent frameworks with fundamental physics and observational data, potentially resolving tensions within the LCDM paradigm. Furthermore, the use of simulated LSST data highlights the survey's potential in offering significant advantages for exploring alternative cosmologies, suggesting that future LSST observations would play a crucial role.

astro-ph.CO↗

arXiv2024

Role of Future SNIa Data from Rubin LSST in Reinvestigating Cosmological Models

We study how future Type-Ia supernovae (SNIa) standard candles detected by the Vera C. Rubin Observatory (LSST) can constrain some cosmological models. We use a realistic three-year SNIa simulated dataset generated by the LSST Dark Energy Science Collaboration (DESC) Time Domain pipeline, which includes a mix of spectroscopic and photometrically identified candidates. We combine this data with Cosmic Microwave Background (CMB) and Baryon Acoustic Oscillation (BAO) measurements to estimate the dark energy model parameters for two models -- the baseline $Λ$CDM and Chevallier-Polarski-Linder (CPL) dark energy parametrization. We compare them with the current constraints obtained from joint analysis of the latest real data from the Pantheon SNIa compilation, CMB from Planck 2018 and BAO. Our analysis finds tighter constraints on the model parameters along with a significant reduction of correlation between $H_0$ and $σ_{8,0}$. We find that LSST is expected to significantly improve upon the existing SNIa data in the critical analysis of cosmological models.

astro-ph.CO↗

arXiv2024

Probing nuclear physics with supernova gravitational waves and machine learning

Core-collapse supernovae are sources of powerful gravitational waves (GWs). We assess the possibility of extracting information about the equation of state (EOS) of high density matter from the GW signal. We use the bounce and early post-bounce signals of rapidly rotating supernovae. A large set of GW signals is generated using general relativistic hydrodynamics simulations for various EOS models. The uncertainty in the electron capture rate is parametrized by generating signals for six different models. To classify EOSs based on the GW data, we train a convolutional neural network (CNN) model. Even with the uncertainty in the electron capture rates, we find that the CNN models can classify the EOSs with an average accuracy of about 87 percent for a set of four distinct EOS models.

astro-ph.HE↗

arXiv2023

Exploring Supernova Gravitational Waves with Machine Learning

Core-collapse supernovae (CCSNe) emit powerful gravitational waves (GWs). Since GWs emitted by a source contain information about the source, observing GWs from CCSNe may allow us to learn more about CCSNs. We study if it is possible to infer the iron core mass from the bounce and early ring-down GW signal. We generate GW signals for a range of stellar models using numerical simulations and apply machine learning to train and classify the signals. We consider an idealized favourable scenario. First, we use rapidly rotating models, which produce stronger GWs than slowly rotating models. Second, we limit ourselves to models with four different masses, which simplifies the selection process. We show that the classification accuracy does not exceed ~70%, signifying that even in this optimistic scenario, the information contained in the bounce and early ring-down GW signal is not sufficient to precisely probe the iron core mass. This suggests that it may be necessary to incorporate additional information such as the GWs from later post-bounce evolution and neutrino observations to accurately measure the iron core mass.

astro-ph.HE↗

arXiv2022

Using Host Galaxy Photometric Redshifts to Improve Cosmological Constraints with Type Ia Supernova in the LSST Era

We perform a rigorous cosmology analysis on simulated type Ia supernovae (SN~Ia) and evaluate the improvement from including photometric host-galaxy redshifts compared to using only the "zspec" subset with spectroscopic redshifts from the host or SN. We use the Deep Drilling Fields (~50 deg^2) from the Photometric LSST Astronomical Time-Series Classification Challenge (PLaSTiCC), in combination with a low-z sample based on Data Challenge2 (DC2). The analysis includes light curve fitting to standardize the SN brightness, a high-statistics simulation to obtain a bias-corrected Hubble diagram, a statistical+systematics covariance matrix including calibration and photo-z uncertainties, and cosmology fitting with a prior from the cosmic microwave background. Compared to using the zspec subset, including events with SN+host photo-z results in i) more precise distances for z>0.5, ii) a Hubble diagram that extends 0.3 further in redshift, and iii) a 50 % increase in the Dark Energy Task Force figure of merit (FoM) based on the w0-wa CDM model. Analyzing 25 simulated data samples, the average bias on w0 and wa is consistent with zero. The host photo-z systematic of 0.01 reduces FoM by only 2 % because i) most z<0.5 events are in the zspec subset, ii) the combined SN+host photo-z has X 2 smaller bias, and iii) the anti-correlation between fitted redshift and color self corrects distance errors. To prepare for analysing real data, the next SNIa-cosmology analysis with photo-z's should include non SN-Ia contamination and host galaxy mis-associations.

astro-ph.CO↗

arXiv2022

Swiss-cheese cosmologies with variable $G$ and $Λ$ from the renormalization group

A convincing explanation for the nature of the dark energy and dark matter is still missing. In recent works a RG-improved swiss-cheese cosmology with an evolving cosmological constant dependent on the \sch radius has been proven to be a promising model to explain the observed cosmic acceleration. In this work we extend this model to consider the combined scaling of the Newton constant $G$ and the cosmological constant $Λ$ according to the IR-fixed point hypothesis. We shall show that our model easily generates the observed recent passage from deceleration to acceleration without need of extra energy scales, exotic fields or fine tuning. In order to check the generality of the concept, two different scaling relations have been analysed and we proved that both are in very good agreement with $Λ$CDM cosmology. We also show that our model satisfies the observational local constraints on $\dot{G}/G$.

gr-qc↗

arXiv2021

SN 2018agk: A Prototypical Type Ia Supernova with a Smooth Power-law Rise in Kepler (K2)

We present the 30-min cadence Kepler/K2 light curve of the Type Ia supernova (SN Ia) SN 2018agk, covering approximately one week before explosion, the full rise phase and the decline until 40 days after peak. We additionally present ground-based observations in multiple bands within the same time range, including the 1-day cadence DECam observations within the first $\sim$5 days after the first light. The Kepler early light curve is fully consistent with a single power-law rise, without evidence of any bump feature. We compare SN 2018agk with a sample of other SNe~Ia without early excess flux from the literature. We find that SNe Ia without excess flux have slowly-evolving early colors in a narrow range ($g-i\approx -0.20\pm0.20$ mag) within the first $\sim 10$ days. On the other hand, among SNe Ia detected with excess, SN 2017cbv and SN 2018oh tend to be bluer, while iPTF16abc's evolution is similar to normal SNe Ia without excess in $g-i$. We further compare the Kepler light curve of SN 2018agk with companion-interaction models, and rule out the existence of a typical non-degenerate companion undergoing Roche-lobe overflow at viewing angles smaller than $45^{\circ}$.

astro-ph.HE↗