arXiv subjects

David Yallup

Publications and source records attributed to David Yallup.

At least 19 recordsLinked to original sources

arXiv2026

Ab Initio Real-Time Gravitational-Wave Parameter Estimation

We present a specialised GPU-native nested sampling kernel targeting rapid parameter estimation for gravitational wave inference problems. Building upon a Slice-within-Gibbs (SwiG) structure for rapid mixing, we investigate how far we can push baseline stochastic sampling techniques on modern GPU hardware. We demonstrate that for typical long-duration binary neutron star signals observed by the LIGO and Virgo detectors, we can achieve well calibrated posterior inference on the full uncompressed data of a three detector network in a median of twelve minutes on a single GPU. This falls to five minutes when sharded across four devices. Utilising heterodyning to compress the data reduces the median wall time across an injection campaign to 89 seconds -- less than the length of the segment itself -- and enables inference with precessing spin, tidal waveforms on GW170817 in around two minutes. This pushes stochastic sampling techniques using full physical waveform calculations, launched from an uninformed prior state, towards real-time gravitational wave parameter estimation.

gr-qc↗

arXiv2026

Rapid Hubble constant inference from GW170817 using GPU-accelerated nested sampling: prior sensitivity and the limits of post-hoc reweighting

The bright-siren measurement of the Hubble constant from GW170817 (Abbott et al. 2017) assumes that switching from a volumetric to a uniform-in-$d_L$ luminosity-distance prior can be implemented by post-hoc reweighting of the baseline samples, rather than by re-running the inference under the target prior. Using a GPU-native heterodyned nested sampling pipeline that completes the full $n_{\rm live}=5000$ analysis in about 13 min on a single A100, we recompute the GW170817 $H_0$ posterior under four prior variants for the modern aligned-spin tidal waveform IMRPhenomXAS_NRTidalv3. Switching from the volumetric to a uniform-in-$d_L$ distance prior raises the high-tail probability $P(H_0>120\,\mathrm{km/s/Mpc})$ from 0.017 to 0.159 when imposed during sampling and shifts the weighted-median $H_0$ from 77.6 to 87.6 km/s/Mpc, while the binned MAP stays at 70.5 km/s/Mpc: both the tail and the bulk move under a change of prior that leaves the mode in place. Post-hoc reweighting of the baseline samples to the same target prior recovers only $P=0.041$ in the tail, approximately 17% of the directly sampled shift. The three prior variants that carry an independent nested sampling evidence agree to $\Delta\ln Z\lesssim 1.8$, so the data show at most a weak preference among the distance priors; the tail and bulk shifts are therefore properties of the prior, not a data update. Targeted mode-isolated runs reveal a $(d_L,\iota)$ bimodality whose high-$H_0$, low-$d_L$ branch (Mode B; $|\ln\mathcal{B}_{\rm B/A}|<1$) the volumetric prior assigns negligible mass: this is the mechanism behind the reweighting deficit. The reweighted posterior has a lower effective sample size than the baseline, independently flagging the coverage failure. The runtime budget makes full-sample prior-sensitivity reruns the default robustness tool for bright-siren cosmology, replacing post-hoc reweighting.

astro-ph.CO↗

arXiv2026

Automatic Laplace Collapsed Sampling: Scalable Marginalisation of Latent Parameters via Automatic Differentiation

We present Automatic Laplace Collapsed Sampling (ALCS), a general framework for marginalising latent parameters in Bayesian models using automatic differentiation, which we combine with nested sampling to explore the hyperparameter space in a robust and efficient manner. At each nested sampling likelihood evaluation, ALCS collapses the high-dimensional latent variables $z$ to a scalar contribution via maximum a posteriori (MAP) optimisation and a Laplace approximation, both computed using autodiff. This reduces the effective dimension from $d_\theta + d_z$ to just $d_\theta$, making Bayesian evidence computation tractable for high-dimensional settings without hand-derived gradients or Hessians, and with minimal model-specific engineering. The MAP optimisation and Hessian evaluation are parallelised across live points on GPU-hardware, making the method practical at scale. We also show that automatic differentiation enables local approximations beyond Laplace to parametric families such as the Student-$t$, which improves evidence estimates for heavy-tailed latents. We validate ALCS on a suite of benchmarks spanning hierarchical, time-series, and discrete-likelihood models and establish where the Gaussian approximation holds. This enables a post-hoc ESS diagnostic that localises failures across hyperparameter space without expensive joint sampling.

cs.LG↗

arXiv2026

The Bayesian view of DESI DR2 with unimpeded: Evidence and tension in a combined analysis with CMB and supernovae across cosmological models

We apply the $\texttt{unimpeded}$ framework to perform a fully Bayesian reanalysis of the DESI DR2 data, using nested sampling with $\texttt{PolyChord}$ to compute evidences for $\Lambda$CDM and seven extensions across combinations of DESI DR1/DR2, Planck CMB, supernovae (Pantheon+, Union3, DES-SN5YR, DES-Dovekie), and DES-Y1 weak lensing. The Bayesian Ockham's razor penalises extended models, yielding weaker or opposite preferences compared to $\Delta\chi^2$-based analyses. For DESI DR2 BAO combined with Planck CMB alone, the DESI collaboration's $3.1\sigma$ frequentist preference for $w_0w_a$CDM is eliminated entirely: we obtain ${\ln B = -0.57{\scriptstyle\pm0.26}}$, modestly favouring $\Lambda$CDM. Adding DES-Dovekie, the recalibration of DES-SN5YR, maintains this concordance (${\ln B = -0.30{\scriptstyle\pm0.19}}$). However, when the earlier DES-SN5YR calibration is included instead, the DESI collaboration's $4.2\sigma$ result survives the Bayesian Ockham penalty as a $3.07{\scriptstyle\pm0.10},\sigma$ preference (${\ln B = +3.32{\scriptstyle\pm0.27}}$). That this signal persists despite the Ockham penalty makes the role of tension quantification essential: our analysis traced the preference to the earlier DES-SN5YR calibration, which introduced a $2.95{\scriptstyle\pm 0.04},\sigma$ conflict with DESI DR2 within $\Lambda$CDM -- a tension that stands out from the grid -- reduced to $1.96{\scriptstyle\pm 0.04},\sigma$ with the DES-Dovekie recalibration. With DES-Dovekie, the Bayesian evidence for dynamical dark energy vanishes.

astro-ph.CO↗

arXiv2026

Nested Sampling with Slice-within-Gibbs: Efficient Evidence Calculation for Hierarchical Bayesian Models

We present Nested Sampling with Slice-within-Gibbs (NS-SwiG), an algorithm for Bayesian inference and evidence estimation in high-dimensional models whose likelihood admits a factorization, such as hierarchical Bayesian models. We construct a procedure to sample from the likelihood-constrained prior using a Slice-within-Gibbs kernel: an outer update of hyperparameters followed by inner block updates over local parameters. A likelihood-budget decomposition caches per-block contributions so that each local update checks feasibility in constant time rather than recomputing the global constraint at linearly growing cost. This reduces the per-replacement cost from quadratic to linear in the number of groups, and the overall algorithmic complexity from cubic to quadratic under standard assumptions. The decomposition extends naturally beyond independent observations, and we demonstrate this on Markov-structured latent variables. We evaluate NS-SwiG on challenging benchmarks, demonstrating scalability to thousands of dimensions and accurate evidence estimates even on posterior geometries where state-of-the-art gradient-based samplers can struggle.

stat.CO↗

arXiv2026

Time delays and stationarity in quasar light curves

We present a fully Bayesian framework for time delay inference and stationarity tests in quasar light curves using marginalised Gaussian processes. The model separates a deterministic, non-stationary drift (piecewise linear mean) from stationary stochastic variability (Mat\'ern and Spectral Mixture kernels), and jointly models multiple images with per-image microlensing. Bayesian evidence and parameter posteriors are obtained via nested sampling and marginalised over model choices. Applied to the quasars WFI J2033 - 4723, B 1608 + 656, and HE 0435 - 1223, we find strong evidence for non-stationarity in B 1608 + 656 and HE 0435 - 1223, while WFI J2033 - 4723 is consistent with stationarity. The stochastic component favours an Markovian exponential kernel for B 1608 + 656 and a non-Markovian Mat\'ern-$\frac32$ kernel for WFI J2033 - 4723 and HE 0435 - 1223. Multi-length-scale Spectral Mixture kernels are disfavoured. Time delays are shown to be robust to model assumptions and consistent with prior work within the error. We further identify and mitigate a likelihood pathology which biases toward large delays, providing a practical nested sampling convergence protocol.

astro-ph.CO↗

arXiv2026

Nested Slice Sampling: Vectorized Nested Sampling for GPU-Accelerated Inference

Model comparison and calibrated uncertainty quantification often require integrating over parameters, but scalable inference can be challenging for complex, multimodal targets. Nested Sampling is a robust alternative to standard MCMC, yet its typically sequential structure and hard constraints make efficient accelerator implementations difficult. This paper introduces Nested Slice Sampling (NSS), a GPU-friendly, vectorized formulation of Nested Sampling that uses Hit-and-Run Slice Sampling for constrained updates. A tuning analysis yields a simple near-optimal rule for setting the slice width, improving high-dimensional behavior and making per-step compute more predictable for parallel execution. Experiments on challenging synthetic targets, high dimensional Bayesian inference, and Gaussian process hyperparameter marginalization show that NSS maintains accurate evidence estimates and high-quality posterior samples, and is particularly robust on difficult multimodal problems where current state-of-the-art methods such as tempered SMC baselines can struggle. An open-source implementation is released to facilitate adoption and reproducibility.

stat.CO↗

arXiv2025

Particle Monte Carlo methods for Lattice Field Theory

High-dimensional multimodal sampling problems from lattice field theory (LFT) have become important benchmarks for machine learning assisted sampling methods. We show that GPU-accelerated particle methods, Sequential Monte Carlo (SMC) and nested sampling, provide a strong classical baseline that matches or outperforms state-of-the-art neural samplers in sample quality and wall-clock time on standard scalar field theory benchmarks, while also estimating the partition function. Using only a single data-driven covariance for tuning, these methods achieve competitive performance without problem-specific structure, raising the bar for when learned proposals justify their training cost.

stat.ML↗

arXiv2025

A Bayesian Perspective on Evidence for Evolving Dark Energy

The DESI Collaboration reports a significant preference for a dynamic dark energy model ($w_0w_a$CDM) over the cosmological constant ($\Lambda$CDM) when their data are combined with other frontier cosmological probes. We present a direct Bayesian model comparison using nested sampling to compute the Bayesian evidence, revealing a contrasting conclusion: for the key combination of the DESI DR2 BAO and the Planck CMB data, we find the Bayesian evidence modestly favours $\Lambda$CDM (log-Bayes factor $\ln B = -0.57{\scriptstyle\pm0.26}$), in contrast to the collaboration's 3.1$\sigma$ frequentist significance in favoring $w_0w_a$CDM. Extending this analysis to also combine with the DES-SN5YR supernova catalogue, our Bayesian analysis reaches a significance of $3.07{\scriptstyle\pm0.10}\,\sigma$ in favour of $w_0w_a$CDM. By performing a comprehensive tension analysis, employing five complementary metrics, we pinpoint the origin: a significant ($2.95{\scriptstyle\pm 0.04}\,\sigma$), low-dimensional tension between DESI DR2 and DES-SN5YR that is present only within the $\Lambda$CDM framework. The $w_0w_a$CDM model is preferred precisely because its additional parameters act to resolve this specific dataset conflict. Replacing DES-SN5YR with the recalibrated DES-Dovekie dataset, this tension is reduced and the three-probe Bayesian evidence for $w_0w_a$CDM vanishes ($\ln B = -0.30{\scriptstyle\pm0.19}$). The convergence of our findings with alternative statistical analyses suggests that the preference for dynamic dark energy is primarily driven by the resolution of inter-dataset tensions, warranting a cautious interpretation of its statistical significance.

astro-ph.CO↗

arXiv2025

Parallel Nested Slice Sampling for Gravitational Wave Parameter Estimation

Inferring parameters and testing hypotheses from gravitational wave signals is a computationally intensive task central to modern astrophysics. Nested sampling, a Bayesian inference technique, has become an established standard for this in the field. However, most common implementations lack the ability to fully utilize modern hardware acceleration. In this work, we demonstrate that when nested sampling is reformulated in a natively vectorized form and run on modern GPU hardware, we can perform inference in a fraction of the time of legacy nested sampling implementations whilst preserving the accuracy and robustness of the method. This scalable, GPU-accelerated approach significantly advances nested sampling for future large-scale gravitational-wave analyses.

astro-ph.IM↗

arXiv2025

High-Dimensional Bayesian Model Comparison in Cosmology with GPU-accelerated Nested Sampling and Neural Emulators

We demonstrate a GPU-accelerated nested sampling framework for efficient high-dimensional Bayesian inference in cosmology. Using JAX-based neural emulators and likelihoods for cosmic microwave background and cosmic shear analyses, our approach provides parameter constraints and direct calculation of Bayesian evidence. In the 39-dimensional $\Lambda$CDM vs $w_0w_a$ shear analysis, we produce Bayes factors and a robust error bar in just 2 days on a single A100 GPU, without loss of accuracy. Where CPU-based nested sampling can now be outpaced by methods relying on MCMC sampling and decoupled evidence estimation, we demonstrate that with GPU acceleration nested sampling offers the necessary speed-up to put it on equal computational footing with these methods, especially where reliable model comparison is paramount. We also explore interpolation in the matter power spectrum for cosmic shear analysis, finding a further factor of 4 speed-up with consistent posterior contours and Bayes factor. We put forward both nested and gradient-based sampling as useful tools for the modern cosmologist, where cutting-edge inference pipelines can yield orders of magnitude improvements in computation time.

astro-ph.CO↗

arXiv2025

Gravitational-wave inference at GPU speed: A bilby-like nested sampling kernel within blackjax-ns

We present a GPU-accelerated implementation of the gravitational-wave Bayesian inference pipeline for parameter estimation and model comparison. Specifically, we implement the `acceptance-walk' sampling method, a cornerstone algorithm for gravitational-wave inference within the bilby and dynesty framework. By integrating this trusted kernel with the vectorized blackjax-ns framework, we achieve typical speedups of 20-40x for aligned spin binary black hole analyses, while recovering posteriors and evidences that are statistically identical to the original CPU implementation. This faithful re-implementation of a community-standard algorithm establishes a foundational benchmark for gravitational-wave inference. It quantifies the performance gains attributable solely to the architectural shift to GPUs, creating a vital reference against which future parallel sampling algorithms can be rigorously assessed. This allows for a clear distinction between algorithmic innovation and the inherent speedup from hardware. Our work provides a validated community tool for performing GPU-accelerated nested sampling in gravitational-wave data analyses.

gr-qc↗

arXiv2023

Kernel-, mean- and noise-marginalised Gaussian processes for exoplanet transits and $H_0$ inference

Using a fully Bayesian approach, Gaussian Process regression is extended to include marginalisation over the kernel choice and kernel hyperparameters. In addition, Bayesian model comparison via the evidence enables direct kernel comparison. The calculation of the joint posterior was implemented with a transdimensional sampler which simultaneously samples over the discrete kernel choice and their hyperparameters by embedding these in a higher-dimensional space, from which samples are taken using nested sampling. Kernel recovery and mean function inference were explored on synthetic data from exoplanet transit light curve simulations. Subsequently, the method was extended to marginalisation over mean functions and noise models and applied to the inference of the present-day Hubble parameter, $H_0$, from real measurements of the Hubble parameter as a function of redshift, derived from the cosmologically model-independent cosmic chronometer and $\Lambda$CDM-dependent baryon acoustic oscillation observations. The inferred $H_0$ values from the cosmic chronometers, baryon acoustic oscillations and combined datasets are $H_0= 66 \pm 6\, \mathrm{km}\,\mathrm{s}^{-1}\,\mathrm{Mpc}^{-1}$, $H_0= 67 \pm 10\, \mathrm{km}\,\mathrm{s}^{-1}\,\mathrm{Mpc}^{-1}$ and $H_0= 69 \pm 6\, \mathrm{km}\,\mathrm{s}^{-1}\,\mathrm{Mpc}^{-1}$, respectively. The kernel posterior of the cosmic chronometers dataset prefers a non-stationary linear kernel. Finally, the datasets are shown to be not in tension with $\ln R=12.17\pm 0.02$.

astro-ph.CO↗

arXiv2023

Hunting for bumps in the margins

Data driven modelling is vital to many analyses at collider experiments, however the derived inference of physical properties becomes subject to details of the model fitting procedure. This work brings a principled Bayesian picture, based on the marginal likelihood, of both data modelling and signal extraction to a common collider physics scenario. First the marginal likelihood based method is used to propose a more principled construction of the background process, systematically exploring a variety of candidate shapes. Second the picture is extended to propose the marginal likelihood as a useful tool for anomaly detection challenges in particle physics. This proposal offers insight into both precise background model determination and demonstrates a flexible method to extend signal determination beyond a simple bump hunt.

physics.data-an↗

arXiv2023

Robust Simulation-Based Inference in Cosmology with Bayesian Neural Networks

Simulation-based inference (SBI) is rapidly establishing itself as a standard machine learning technique for analyzing data in cosmological surveys. Despite continual improvements to the quality of density estimation by learned models, applications of such techniques to real data are entirely reliant on the generalization power of neural networks far outside the training distribution, which is mostly unconstrained. Due to the imperfections in scientist-created simulations, and the large computational expense of generating all possible parameter combinations, SBI methods in cosmology are vulnerable to such generalization issues. Here, we discuss the effects of both issues, and show how using a Bayesian neural network framework for training SBI can mitigate biases, and result in more reliable inference outside the training set. We introduce cosmoSWAG, the first application of Stochastic Weight Averaging to cosmology, and apply it to SBI trained for inference on the cosmic microwave background.

astro-ph.CO↗

arXiv2022

Exploring phase space with Nested Sampling

We present the first application of a Nested Sampling algorithm to explore the high-dimensional phase space of particle collision events. We describe the adaptation of the algorithm, designed to perform Bayesian inference computations, to the integration of partonic scattering cross sections and the generation of individual events distributed according to the corresponding squared matrix element. As a first concrete example we consider gluon scattering processes into 3-, 4- and 5-gluon final states and compare the performance with established sampling techniques. Starting from a flat prior distribution Nested Sampling outperforms the Vegas algorithm and achieves results comparable to a dedicated multi-channel importance sampler. We outline possible approaches to combine Nested Sampling with non-flat prior distributions to further reduce the variance of integral estimates and to increase unweighting efficiencies.

hep-ph↗

arXiv2022

Nested sampling for physical scientists

We review Skilling's nested sampling (NS) algorithm for Bayesian inference and more broadly multi-dimensional integration. After recapitulating the principles of NS, we survey developments in implementing efficient NS algorithms in practice in high-dimensions, including methods for sampling from the so-called constrained prior. We outline the ways in which NS may be applied and describe the application of NS in three scientific fields in which the algorithm has proved to be useful: cosmology, gravitational-wave astronomy, and materials science. We close by making recommendations for best practice when using NS and by summarizing potential limitations and optimizations of NS.

stat.CO↗

arXiv2022

Split personalities in Bayesian Neural Networks: the case for full marginalisation

The true posterior distribution of a Bayesian neural network is massively multimodal. Whilst most of these modes are functionally equivalent, we demonstrate that there remains a level of real multimodality that manifests in even the simplest neural network setups. It is only by fully marginalising over all posterior modes, using appropriate Bayesian sampling tools, that we can capture the split personalities of the network. The ability of a network trained in this manner to reason between multiple candidate solutions dramatically improves the generalisability of the model, a feature we contend is not consistently captured by alternative approaches to the training of Bayesian neural networks. We provide a concise minimal example of this, which can provide lessons and a future path forward for correctly utilising the explainability and interpretability of Bayesian neural networks.

stat.ML↗