arXiv subjects

Matthew O'Callaghan

Publications and source records attributed to Matthew O'Callaghan.

5 recordsLinked to original sources

arXiv2026

On the origin of the environmental step: A BayeSN view of the ZTF SN Ia DR2

Astrophysical variabilities of Type Ia supernovae (SNe Ia), such as their link with their birth environment, are now one of the leading sources of systematic uncertainties on the measurement of the dark energy equation-of-state parameter $w$. Population studies of SNe Ia, using large samples, give precious insights into these variabilities. We analyse a volume-limited subsample of the ZTF SN Ia DR2 with BayeSN, a hierarchical Bayesian model for SN Ia SEDs. We investigate the distributions of SN Ia light curve parameters and their link with SN environment. Using a new training of BayeSN released in a companion paper, we find a smaller scatter of Hubble residuals compared to SALT. We then investigate the magnitude step, which accounts for the correlation between SN Ia standardised absolute magnitude and host environments. We find a posteriori steps of $0.103\pm0.010$ mag (a $10.1\sigma$ difference from 0) when using global stellar mass as an environmental proxy, and $0.086\pm0.010$ mag ($8.3\sigma$) when using local colour, in accordance with steps computed using SALT light curve fits. This confirms that the large step seen in the ZTF SN Ia DR2 data was not due to the SALT fit or the associated standardisation process. We then investigate the origin of the step, using a BayeSN model which accounts for both an intrinsic magnitude step and differing dust properties with the SN environment. We find a $0.103\pm0.018$ mag ($5.6\sigma$) step in global mass and a $0.085\pm0.019$ mag ($4.5\sigma$) step in local colour. The means of the $R_V$ distribution are similar between different host environments, with $\Delta\mathbb{E}(R_V)\leq0.2$ across all environment proxies, with significances ranging from $0.6\sigma$ to $1.2\sigma$. This is a strong signal of the existence of an intrinsic dependence of SN Ia absolute magnitude on environment.

astro-ph.CO↗

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↗

arXiv2025

Misspecification-robust amortised simulation-based inference using variational methods

Recent advances in neural density estimation have enabled powerful simulation-based inference (SBI) methods that can flexibly approximate Bayesian inference for intractable stochastic models. Although these methods have demonstrated reliable posterior estimation when the simulator accurately represents the underlying data generative process (DGP), recent work has shown that they perform poorly in the presence of model misspecification. This poses a significant issue for their use in real-world problems, due to simulators always misrepresenting the true DGP to a certain degree. In this paper, we introduce robust variational neural posterior estimation (RVNP), a method which addresses the problem of misspecification in amortised SBI by bridging the simulation-to-reality gap using variational inference and error modelling. We test RVNP on multiple benchmark tasks, including using real data from astronomy, and show that it can recover robust posterior inference in a data-driven manner without adopting hyperparameters or priors governing the misspecification influence.

stat.ML↗

arXiv2025

Data-driven dust inference at mid-to-high Galactic latitudes using probabilistic machine learning

We present a method for accurately and precisely inferring photometric dust extinction towards stars at mid-to-high Galactic latitudes using probabilistic machine learning to model the colour-magnitude distribution of zero-extinction stars in these regions. Photometric dust maps rely on a robust method for inferring stellar reddening. At high Galactic latitudes, where extinction is low, such inferences are particularly susceptible to contamination from modelling errors and prior assumptions, potentially introducing artificial structure into dust maps. In this work, we demonstrate the use of normalising flows to learn the conditional probability distribution of the photometric colour-magnitude relations of zero-extinction stars, conditioned on Galactic cylindrical coordinates for stars at mid-to-high Galactic latitudes. By using the normalising flow to model the colour-magnitude diagram, we infer the posterior distribution of dust extinction towards stars along different lines of sight by marginalising over the colour-magnitude flow. We validate our method using data from Gaia, Pan-STARRS, and 2MASS, showing that we can recover unbiased posteriors and successfully detect dust along the line-of-sight in two calibration regions at mid-Galactic latitude that have been extensively studied in the context of polarisation surveys.

astro-ph.GA↗

arXiv2023

Quantifying Interstellar Extinction at High Galactic Latitudes

A detailed map of the distribution of dust at high Galactic latitudes is essential for future cosmic microwave background (CMB) polarization experiments because the dust, while diffuse, remains a significant foreground in these regions. We develop a Bayesian model to identify a region of the Hertzsprung-Russell (HR) diagram suited to constrain the single-star extinction accurately at high Galactic latitudes. Using photometry from Gaia, 2MASS and ALLWISE, and parallax from Gaia, we employ nested sampling to fit the model to the data and analyse the posterior over stellar parameters for both synthetic and real data. Charting low variations in extinction is complex due to systematic errors and degeneracies between extinction and other stellar parameters. The systematic errors can be minimised by restricting our data to a region of the HR diagram where the stellar models are most accurate. Moreover, the degeneracies can be reduced by including astrophysical priors and spectroscopic constraints. We show that accounting for the measurement error of the data and the assumed inaccuracies of the stellar models are critical in accurately recovering small variations in extinction. We compare the posterior distribution for individual stars with spectroscopic stellar parameter estimates from LAMOST and Gaia ESO and illustrate the importance of estimating extinction and effective temperature as a joint solution.

astro-ph.GA↗