arXiv subjects

Liam Pinchbeck

Publications and source records attributed to Liam Pinchbeck.

3 recordsLinked to original sources

arXiv2026

$\texttt{BilbyFlow}$: user-friendly neural posterior estimation for gravitational-wave astronomy

Bayesian inference plays a central role in the new field of gravitational-wave astronomy. However, traditional Bayesian inference with stochastic samplers is computationally expensive, taking hours to days per event. Transformative changes are therefore required to enable the science of next-generation observatories whose event rates and signal-to-noise ratios will increase significantly over the current generation. Recent work has shown that neural posterior estimation (NPE) is a promising path forward. A neural net is trained to approximate the posterior distribution of gravitational-wave parameters, allowing generation of posterior samples in a fraction of the time required by stochastic samplers. In this work, we introduce $\texttt{BilbyFlow}$, which harnesses the power of NPE in the popular $\texttt{Bilby}$ code suite. We use $\texttt{BilbyFlow}$ to analyze a subset of 38 high-mass events from the third LIGO-Virgo-KAGRA Gravitational-Wave Transient Catalog (GWTC-3). For 29 events (76\%), we obtained an importance-sampling efficiency $>$1%, allowing us to produce reliable posterior distributions within 3 min - 1.5 hours. For the other events, with importance-sampling efficiency $\ll$1%, the run time can be as long as 35 hours. We achieve a median importance-sampling efficiency of 7%, which is roughly comparable to the $\texttt{DINGO}$ package. We aim to significantly improve this efficiency with further development to make the runtime more reliably $O(\text{min})$. $\texttt{BilbyFlow}$ is open source and $\texttt{pip}$-installable.

astro-ph.IM↗

arXiv2025

Model-independent dark matter detection with the Cherenkov Telescope Array Observatory

Searches for annihilating dark matter are often designed with a specific dark matter candidate in mind. However, the space of potential dark matter models is vast, which raises the question: how can we search for dark matter without making strong assumptions about unknown physics. We present a model-independent approach for measuring dark matter annihilation ratios and branching fractions with $γ$-ray event data. By parameterizing the annihilation ratios for seven different channels, we obviate the need to search for a specific dark matter candidate. To demonstrate our approach, we analyse simulated data using the GammaBayes pipeline. Given a 5$σ$ signal, we reconstruct the annihilation ratios for five dominant channels to within 95% credibility. This allows us to reconstruct dark matter annihilation/decay channels without presuming any particular model, thus offering a model-independent approach to indirect dark matter searches in $γ$-ray astronomy. This approach shows that for masses between 0.3-5 TeV we can probe values below the thermal relic velocity annihilation weighted cross-section allowing a 2$σ$ detection for 525 hours of simulated observation data by the Cherenkov Telescope Array Observatory of the Galactic Centre.

astro-ph.HE↗

arXiv2024

GammaBayes: a Bayesian pipeline for dark matter detection with CTA

We present GammaBayes, a Bayesian Python package for dark matter detection with the Cherenkov Telescope Array (CTA). GammaBayes takes as input the CTA measurements of gamma rays and a user-specified dark-matter particle model. It outputs the posterior distribution for parameters of the dark-matter model including the velocity-averaged cross section for dark-matter self interactions $\langleσv\rangle$ and the dark-matter mass $m_χ$. It also outputs the Bayesian evidence, which can be used for model selection. We demonstrate GammaBayes using 525 hours of simulated data, corresponding to $10^8$ observed gamma-ray events. The vast majority of this simulated data consists of noise, but $100000$ events arise from the annihilation of scalar singlet dark matter with $m_χ= 1$ TeV. We recover the dark matter mass within a 95% credible interval of $m_χ\sim 0.96-1.07$ TeV. Meanwhile, the velocity averaged cross section is constrained to $\langleσv\rangle \sim 1.4-2.1\times10^{-25}$ cm$^3$ s$^{-1}$ (95% credibility). This is equivalent to measuring the number of dark-matter annihilation events to be $N_S \sim 1.1_{-0.2}^{+0.2} \times 10^5$. The no-signal hypothesis $\langle σv \rangle=0$ is ruled out with about $5σ$ credibility. We discuss how GammaBayes can be extended to include more sophisticated signal and background models and the computational challenges that must be addressed to facilitate these upgrades. The source code is publicly available at https://github.com/lpin0002/GammaBayes.

astro-ph.HE↗