arXiv subjects

Samuel Powell

Publications and source records attributed to Samuel Powell.

7 recordsLinked to original sources

arXiv2026

HealthBench-Psych: A Mental Health Subset of OpenAI's HealthBench

General-purpose health benchmarks increasingly anchor claims about LLM medical performance, but they are not always resolved by clinical specialty, making domain-specific performance hard to isolate. Mental health is of acute public-health concern as millions of people turn to LLMs for psychological support, and most existing evaluations are bespoke academic benchmarks that are difficult to integrate into developer workflows. We introduce HealthBench-Psych and HealthBench-Psych-Hard. We screened HealthBench's 5,000 physician-rubric conversations for mental-health relevance with a transparent LLM-applied rubric, then validated the subset through two rounds of blinded clinician review with concealed known-exclude controls, yielding 610 conversations (12.2% of the corpus). Evaluating 20 frontier and open models under a cross-vendor panel of three LLM judges, we find a statistically tied frontier cluster, measurable refusal behavior in two models, and near-identical rankings across judges ($\tau \ge 0.92$). We release the subset, pipeline, model responses, grades, and analysis code as a reusable resource.

cs.CL↗

arXiv2026

The RD50-MPW4: A Radiation Hard HV CMOS Sensor for Future Colliders

The former CERN RD50 collaboration develops monolithic active pixel high voltage (HV) CMOS sensors for future colliders with the aim of high radiation tolerance, good time resolution, and high granularity pixel detectors. The most recent prototype, the RD50-MPW4, was produced by LFoundry in December 2023 using a 150 nm CMOS process. It features a matrix of 64x64 pixels with a 62 $\mu$m pitch and employs a column-drain readout architecture. Compared to its predecessor, it now has separate analog and digital power domains and a new biasing scheme with a guard ring structure that supports bias voltages over 600 V. This contribution will discuss the design and latest results of the MPW4, where tests with unirradiated samples showed more than 99.9% efficiency, 16 $\mu$m spatial resolution and 10 ns timing resolution. Efficiencies of over 99% were achieved for samples irradiated to fluences of $1 \times 10^{15}$ 1 MeV n$_{\mathrm{eq}}/\mathrm{cm}^2$. A 3D map of the charge collection efficiency from a measurement with two-photon absorption laser is presented that nicely outlines the depletion depth of this radiation hard, high granularity monolithic active pixel sensor.

physics.ins-det↗

arXiv2020

Efficient inversion strategies for estimating optical properties with Monte Carlo radiative transport models

Indirect imaging problems in biomedical optics generally require repeated evaluation of forward models of radiative transport, for which Monte Carlo is accurate yet computationally costly. We develop a novel approach to reduce this bottleneck which has significant implications for quantitative tomographic imaging in a variety of medical and industrial applications. Using Monte Carlo we compute a fully stochastic gradient of an objective function for a given imaging problem. Leveraging techniques from the machine learning community we then adaptively control the accuracy of this gradient throughout the iterative inversion scheme, in order to substantially reduce computational resources at each step. For example problems of Quantitative Photoacoustic Tomography and Ultrasound Modulated Optical Tomography, we demonstrate that solutions are attainable using a total computational expense that is comparable to (or less than) that which is required for a single high accuracy forward run of the same Monte Carlo model. This approach demonstrates significant computational savings when approaching the full non-linear inverse problem of optical property estimation using stochastic methods.

physics.comp-ph↗

arXiv2018

A pseudospectral method for solution of the radiative transport equation

The radiative transport equation accurately describes light transport in participating media such as biological tissues, though analytic solutions are known only for simple geometries. We present a pseudospectral technique to efficiently compute numerical solutions to the time-dependent transport equation in arbitrary three-dimensional heterogeneous media with anisotropic scattering. Our GPU-accelerated implementation of the technique is validated by comparison with a Monte-Carlo simulation, demonstrating excellent agreement.

physics.comp-ph↗

arXiv2016

Quantitative photoacoustic tomography using forward and adjoint Monte Carlo models of radiance

Forward and adjoint Monte Carlo (MC) models of radiance are proposed for use in model-based quantitative photoacoustic tomography. A 2D radiance MC model using a harmonic angular basis is introduced and validated against analytic solutions for the radiance in heterogeneous media. A gradient-based optimisation scheme is then used to recover 2D absorption and scattering coefficients distributions from simulated photoacoustic measurements. It is shown that the functional gradients, which are a challenge to compute efficiently using MC models, can be calculated directly from the coefficients of the harmonic angular basis used in the forward and adjoint models. This work establishes a framework for transport-based quantitative photoacoustic tomography that can fully exploit emerging highly parallel computing architectures.

physics.med-ph↗

arXiv2015

Reconstruction-classification method for quantitative photoacoustic tomography

We propose a combined reconstruction-classification method for simultaneously recovering absorption and scattering in turbid media from images of absorbed optical energy. This method exploits knowledge that optical parameters are determined by a limited number of classes to iteratively improve their estimate. Numerical experiments show that the proposed approach allows for accurate recovery of absorption and scattering in 2 and 3 dimensions, and delivers superior image quality with respect to traditional reconstruction-only approaches.

physics.med-ph↗

arXiv2014

Gradient-based quantitative image reconstruction in ultrasound-modulated optical tomography: first harmonic measurement type in a linearised diffusion formulation

Ultrasound-modulated optical tomography is an emerging biomedical imaging modality which uses the spatially localised acoustically-driven modulation of coherent light as a probe of the structure and optical properties of biological tissues. In this work we begin by providing an overview of forward modelling methods, before deriving a linearised diffusion-style model which calculates the first-harmonic modulated flux measured on the boundary of a given domain. We derive and examine the correlation measurement density functions of the model which describe the sensitivity of the modality to perturbations in the optical parameters of interest. Finally, we employ said functions in the development of an adjoint-assisted gradient based image reconstruction method, which ameliorates the computational burden and memory requirements of a traditional Newton-based optimisation approach. We validate our work by performing reconstructions of optical absorption and scattering in two- and three-dimensions using simulated measurements with 1% proportional Gaussian noise, and demonstrate the successful recovery of the parameters to within +/-5% of their true values when the resolution of the ultrasound raster probing the domain is sufficient to delineate perturbing inclusions.

physics.med-ph↗