arXiv subjects

Leigh H. Whitehead

Publications and source records attributed to Leigh H. Whitehead.

13 recordsLinked to original sources

arXiv2026

Physics at the Edge: Benchmarking Quantisation Techniques and the Edge TPU for Neutrino Interaction Recognition

This work presents a benchmark of quantisation techniques for convolutional neural networks applied to neutrino interaction recognition, motivated by the need for low-power inference close to the detector in future liquid argon time-projection chamber (LArTPC) experiments. Using simulation of a generic LArTPC, four Keras models are trained to classify neutral-current, muon-neutrino charged-current and electron-neutrino charged-current interactions, quantised using post-training integer quantisation and quantisation-aware training, and deployed on the Google Coral Edge TPU. Latency and energy per event are compared with those of an Intel Core i7-13650HX CPU, a consumer NVIDIA RTX 4060 GPU and a data-centre NVIDIA A100 GPU, using measured energy at different batch sizes and thread counts; the energy of the Edge TPU is estimated from its nominal power. The accuracy degradation is strongly model-dependent: InceptionV3 shows no significant loss along either pipeline, whereas the other models are substantially degraded. At batch size 1, the Edge TPU is comparable to or faster than the CPU and 1.3-6.5 times slower than the A100, while requiring 5--70 times less energy per event than the other devices; with batching, the GPUs reduce this gap to a factor of 2.6-7.2. The Edge TPU is also approximately four times cheaper than the consumer GPU and more than two orders of magnitude cheaper than the A100. Its low power per device eases deployment where power and cooling are constrained, with an additional benefit in terms of environmental impact. This paper explores possible future integrations of edge AI technologies with neutrino physics.

physics.ins-det↗

arXiv2026

Submanifold Sparse Convolutional Networks for Automated 3D Segmentation of Kidneys and Kidney Tumours in Computed Tomography

Accurate delineation of kidney tumours in Computed Tomography (CT) is essential for downstream quantitative analysis and precision oncology, but manual segmentation is a specialised task, time-consuming and difficult to scale. Automated 3D segmentation remains challenging because CT scans are large volumetric images, making high-resolution dense convolutional networks computationally expensive and often dependent on downsampling or patch-based inference. We propose a two-stage 3D segmentation methodology based on voxel sparsification and submanifold sparse convolutional networks (SSCNs). Stage 1 uses a low-resolution sparse network to identify a region of interest (ROI); Stage 2 applies a high-resolution sparse network for refined segmentation within the cropped ROI. This enables native high-resolution 3D processing while reducing memory use and inference time. We evaluate the method on the KiTS23 renal cancer CT dataset using 5-fold cross-validation. Our method achieved Dice similarity coefficients of 95.8% for kidneys + masses, 85.7% for tumours + cysts, and 80.3% for tumours alone, competitive with top KiTS23 approaches. In direct comparisons on the same cross-validation folds, the proposed sparse method achieves tumour + cyst and tumour-only Dice scores comparable to, and slightly higher than, a patch-based nnU-Net baseline, while consistently requiring less VRAM and shorter inference time across the tested hardware. Across the tested GPUs, our sparse model is markedly faster than both nnU-Net and the zero-shot zoom-out/zoom-in foundation model SegVol, which localises kidneys well but underperforms on small heterogeneous lesions. Compared to an equivalent dense implementation of the same architecture, the proposed sparse approach achieves up to a 60% reduction in inference time and up to a 75% reduction in VRAM usage across both CPU and the GPU configurations tested.

cs.CV↗

arXiv2025

LArTPC hit-based topology classification with quantum machine learning and symmetry

We present a new approach to separate track-like and shower-like topologies in liquid argon time projection chamber (LArTPC) experiments for neutrino physics using quantum machine learning. Effective reconstruction of neutrino events in LArTPCs requires accurate and granular information about the energy deposited in the detector. These energy deposits can be viewed as 2-D images. Simulated data from the MicroBooNE experiment and a simple custom dataset are used to perform pixel-level classification of the underlying particle topology. Images of the events have been studied by creating small patches around each pixel to characterise its topology based on its immediate neighbourhood. This classification is achieved using convolution-based learning models, including quantum-enhanced architectures known as quanvolutional neural networks. The quanvolutional networks are extended to symmetries beyond translation. Rotational symmetry has been incorporated into a subset of the models. Quantum-enhanced models perform better than their classical counterparts with a comparable number of parameters but are outperformed by classical models, which contain an order of magnitude more parameters. The inclusion of rotation symmetry appears to benefit only large models and remains to be explored further.

physics.ins-det↗

arXiv2024

Scalable DAQ system operating the CHIPS-5 neutrino detector

The CHIPS R&D project focuses on development of low-cost water Cherenkov neutrino detectors through novel design strategies and resourceful engineering. This work presents an end-to-end DAQ solution intended for a recent 5 kt CHIPS prototype, which is largely based on affordable mass-produced components. Much like the detector itself, the presented instrumentation is composed of modular arrays that can be scaled up and easily serviced. A single such array can carry up to 30 photomultiplier tubes (PMTs) accompanied by electronics that generate high voltage in-situ and deliver time resolution of up to 0.69 ns. In addition, the technology is compatible with the White Rabbit timing system, which can synchronize its elements to within 100 ps. While deployment issues did not permit the presented DAQ system to operate beyond initial evaluation, the presented hardware and software successfully passed numerous commissioning tests that demonstrated their viability for use in a large-scale neutrino detector, instrumented with thousands of PMTs.

physics.ins-det↗

arXiv2023

Application of Transfer Learning to Neutrino Interaction Classification

Training deep neural networks using simulations typically requires very large numbers of simulated events. This can be a large computational burden and a limitation in the performance of the deep learning algorithm when insufficient numbers of events can be produced. We investigate the use of transfer learning, where a set of simulated images are used to fine tune a model trained on generic image recognition tasks, to the specific use case of neutrino interaction classification in a liquid argon time projection chamber. A ResNet18, pre-trained on photographic images, was fine-tuned using simulated neutrino images and when trained with one hundred thousand training events reached an F1 score of $0.896 \pm 0.002$ compared to $0.836 \pm 0.004$ from a randomly-initialised network trained with the same training sample. The transfer-learned networks also demonstrate lower bias as a function of energy and more balanced performance across different interaction types.

hep-ex↗

arXiv2022

Automated Segmentation of Computed Tomography Images with Submanifold Sparse Convolutional Networks

Quantitative cancer image analysis relies on the accurate delineation of tumours, a very specialised and time-consuming task. For this reason, methods for automated segmentation of tumours in medical imaging have been extensively developed in recent years, being Computed Tomography one of the most popular imaging modalities explored. However, the large amount of 3D voxels in a typical scan is prohibitive for the entire volume to be analysed at once in conventional hardware. To overcome this issue, the processes of downsampling and/or resampling are generally implemented when using traditional convolutional neural networks in medical imaging. In this paper, we propose a new methodology that introduces a process of sparsification of the input images and submanifold sparse convolutional networks as an alternative to downsampling. As a proof of concept, we applied this new methodology to Computed Tomography images of renal cancer patients, obtaining performances of segmentations of kidneys and tumours competitive with previous methods (~84.6% Dice similarity coefficient), while achieving a significant improvement in computation time (2-3 min per training epoch).

eess.IV↗

arXiv2022

End-to-end analysis using image classification

End-to-end analyses of data from high-energy physics experiments using machine and deep learning techniques have emerged in recent years. These analyses use deep learning algorithms to go directly from low-level detector information directly to high-level quantities that classify the interactions. The most popular class of algorithms for these analyses are convolutional neural networks that operate on experimental data formatted as images. End-to-end analyses skip stages of the traditional workflow that includes the reconstruction of particles produced in the interactions, and as such are not limited by efficiency losses and sources of inaccuracy throughout the event reconstruction process. In many cases, deep learning end-to-end analyses have been shown to have significantly increased performance compared to previous state-of-the-art methods.

physics.data-an↗

arXiv2022

Detector and Beamline Simulation for Next-Generation High Energy Physics Experiments

The success of high energy physics programs relies heavily on accurate detector simulations and beam interaction modeling. The increasingly complex detector geometries and beam dynamics require sophisticated techniques in order to meet the demands of current and future experiments. Common software tools used today are unable to fully utilize modern computational resources, while data-recording rates are often orders of magnitude larger than what can be produced via simulation. In this paper, we describe the state, current and future needs of high energy physics detector and beamline simulations and related challenges, and we propose a number of possible ways to address them.

hep-ex↗

arXiv2021

Graph neural network for 3D classification of ambiguities and optical crosstalk in scintillator-based neutrino detectors

Deep learning tools are being used extensively in high energy physics and are becoming central in the reconstruction of neutrino interactions in particle detectors. In this work, we report on the performance of a graph neural network in assisting with particle flow event reconstruction. The three-dimensional reconstruction of particle tracks produced in neutrino interactions can be subject to ambiguities due to high multiplicity signatures in the detector or leakage of signal between neighboring active detector volumes. Graph neural networks potentially have the capability of identifying all these features to boost the reconstruction performance. As an example case study, we tested a graph neural network, inspired by the GraphSAGE algorithm, on a novel 3D-granular plastic-scintillator detector, that will be used to upgrade the near detector of the T2K experiment. The developed neural network has been trained and tested on diverse neutrino interaction samples, showing very promising results: the classification of particle track voxels produced in the detector can be done with efficiencies and purities of 94-96% per event and most of the ambiguities can be identified and rejected, while being robust against systematic effects.

physics.data-an↗

arXiv2021

Probing Energetic Light Dark Matter with Multi-Particle Tracks Signatures at DUNE

The search for relativistic scattering signals of cosmogenic light dark matter at terrestrial detectors has received increasing attention as an alternative approach to probe dark-sector physics. Large-volume neutrino experiments are well motivated for searches of dark matter that interacts very weakly with Standard Model particles and/or that exhibits a small incoming flux. We perform a dedicated signal sensitivity study for a detector similar to the one proposed by the DUNE Collaboration for cosmogenic dark-matter signals resulting from a non-minimal multi-particle dark-sector scenario. The liquid argon time projection chamber technology adopted for the DUNE detectors is particularly suited for searching for complicated signatures owing to good measurement resolution and particle identification, as well as $dE/dx$ measurements to recognize merged tracks. Taking inelastic boosted dark matter as our benchmark scenario that allows for multiple visible particles in the final state, we demonstrate that the DUNE far detectors have a great potential for probing scattering signals induced by relativistic light dark matter. Detector effects and backgrounds have been estimated and taken into account. Model-dependent and model-independent expected sensitivity limits for a DUNE-like detector are presented.

hep-ph↗

arXiv2020

Image-based model parameter optimization using Model-Assisted Generative Adversarial Networks

We propose and demonstrate the use of a model-assisted generative adversarial network (GAN) to produce fake images that accurately match true images through the variation of the parameters of the model that describes the features of the images. The generator learns the model parameter values that produce fake images that best match the true images. Two case studies show excellent agreement between the generated best match parameters and the true parameters. The best match model parameter values can be used to retune the default simulation to minimize any bias when applying image recognition techniques to fake and true images. In the case of a real-world experiment, the true images are experimental data with unknown true model parameter values, and the fake images are produced by a simulation that takes the model parameters as input. The model-assisted GAN uses a convolutional neural network to emulate the simulation for all parameter values that, when trained, can be used as a conditional generator for fast fake-image production.

cs.CV↗

arXiv2019

Search for Boosted Dark Matter at ProtoDUNE

We propose the first experimental test of the inelastic boosted dark matter hypothesis, capitalizing on the new physics potential with the imminent data taking of the ProtoDUNE detectors. More specifically, we explore various experimental signatures at the cosmic frontier, arising in boosted dark matter scenarios, i.e., relativistic, inelastic scattering of boosted dark matter often created by the annihilation of its heavier component which usually comprises of the dominant relic abundance. Although features are unique enough to isolate signal events from potential backgrounds, vetoing a vast amount of cosmic background is rather challenging as the detectors are located on the ground. We argue, with a careful estimate, that such backgrounds nevertheless can be well under control by performing dedicated analyses after data acquisition. We then discuss some phenomenological studies which can be achieved with ProtoDUNE, employing a dark photon scenario as our benchmark dark-sector model.

hep-ph↗

arXiv2016

Neutrino Oscillations with MINOS and MINOS+

The MINOS experiment ran from 2003 until 2012 and collected a data sample including 10.71x10^20 protons-on-target (POT) of beam neutrinos, 3.36x10^20 POT of beam antineutrinos and an atmospheric neutrino exposure of 37.88 kt-yrs. The final measurement of the atmospheric neutrino oscillation parameters, dm^2_32 and theta_23, came from a full three flavour oscillation analysis of the combined CC nu_mu and CC anti-nu_mu beam and atmospheric samples and the CC nu_e and CC anti-nu_e appearance samples. This analysis yielded the most precise measurement of the atmospheric mass splitting dm^2_32 performed to date. The results are |dm^2_32|=[2.28 - 2.46]x10^-3 eV^2 (68\%) and sin^{2}theta_23=0.35-0.65$ (90\%) in the normal hierarchy, and |dm^2_32|=[2.32 - 2.53]x10^-3 eV^2 (68\%) and sin^{2}theta_23=0.34-0.67 (90\%) in the inverted hierarchy. The successor to MINOS in the NOvA era at FNAL, MINOS+, is now collecting data mostly in the 3-10 GeV region, and an analysis of nu_mu disappearance using the first 2.99x10^20 POT of data produced results very consistent with those from MINOS. Future data will further test the standard neutrino oscillation paradigm and allow for improved searches for exotic phenomena including sterile neutrinos, large extra dimensions and non-standard interactions.

hep-ex↗