arXiv subjects

Grant Stevens

Publications and source records attributed to Grant Stevens.

8 recordsLinked to original sources

arXiv2026

DeepDISC-Euclid: Source Classification and Photometric Redshifts in Euclid Deep Field North With a Pixel-Level Deep Learning Approach

The first Euclid Quick Data Release (Q1) provides extensive imaging and spectroscopic data for hundreds of millions of photometric objects across several deep fields. Accurate classifications and photometric redshifts (photo-z) for these sources are crucial to maximizing the value of these data. In this work, we perform source classification and photo-z estimation for the Euclid Deep Field North (EDF-N) around the North Ecliptic Pole, using a deep learning framework (DeepDISC) that learns and infers using 9-band images simultaneously. We train three dedicated models for (1) source detection and classification, (2) galaxy photo-z, and (3) quasar photo-z. The Euclid Q1 input source catalog, and classifications and spectroscopic redshifts (spec-z) from the Dark Energy Spectroscopic Instrument Data Release 1 are adopted as our training data. DeepDISC source detection achieves overall completeness of ~93% and purity of ~80% if using the Euclid source catalog as the ground truth. Using a JWST source catalog within EDF-N as the reference, we estimate a true purity of ~ 90% for DeepDISC sources. About 99.2%, 99.0%, and 84.8% of stars, galaxies, and quasars, respectively, are correctly recovered with their spectroscopic classifications. The DeepDISC photo-zs show good agreement with spectroscopic redshifts, for both galaxies and quasars. Comparisons with other Euclid Q1 products demonstrate that DeepDISC provides comparable or improved performance in source detection/deblending, classification and photo-z, especially for quasars. These results demonstrate the potential of pixel-level deep learning approaches for large-scale sky surveys such as Euclid and Roman, which will continue to improve with better training labels. We release the full DeepDISC source catalog (~13 million objects) for EDF-N with classifications and photo-zs, including photo-z probability distributions.

astro-ph.GA↗

arXiv2025

Scout-Dose-TCM: Direct and Prospective Scout-Based Estimation of Personalized Organ Doses from Tube Current Modulated CT Exams

This study proposes Scout-Dose-TCM for direct, prospective estimation of organ-level doses under tube current modulation (TCM) and compares its performance to two established methods. We analyzed contrast-enhanced chest-abdomen-pelvis CT scans from 130 adults (120 kVp, TCM). Reference doses for six organs (lungs, kidneys, liver, pancreas, bladder, spleen) were calculated using MC-GPU and TotalSegmentator. Based on these, we trained Scout-Dose-TCM, a deep learning model that predicts organ doses corresponding to discrete cosine transform (DCT) basis functions, enabling real-time estimates for any TCM profile. The model combines a feature learning module that extracts contextual information from lateral and frontal scouts and scan range with a dose learning module that output DCT-based dose estimates. A customized loss function incorporated the DCT formulation during training. For comparison, we implemented size-specific dose estimation per AAPM TG 204 (Global CTDIvol) and its organ-level TCM-adapted version (Organ CTDIvol). A 5-fold cross-validation assessed generalizability by comparing mean absolute percentage dose errors and r-squared correlations with benchmark doses. Average absolute percentage errors were 13% (Global CTDIvol), 9% (Organ CTDIvol), and 7% (Scout-Dose-TCM), with bladder showing the largest discrepancies (15%, 13%, and 9%). Statistical tests confirmed Scout-Dose-TCM significantly reduced errors vs. Global CTDIvol across most organs and improved over Organ CTDIvol for the liver, bladder, and pancreas. It also achieved higher r-squared values, indicating stronger agreement with Monte Carlo benchmarks. Scout-Dose-TCM outperformed Global CTDIvol and was comparable to or better than Organ CTDIvol, without requiring organ segmentations at inference, demonstrating its promise as a tool for prospective organ-level dose estimation in CT.

physics.med-ph↗

arXiv2025

Towards Foundational Models for Dynamical System Reconstruction: Hierarchical Meta-Learning via Mixture of Experts

As foundational models reshape scientific discovery, a bottleneck persists in dynamical system reconstruction (DSR): the ability to learn across system hierarchies. Many meta-learning approaches have been applied successfully to single systems, but falter when confronted with sparse, loosely related datasets requiring multiple hierarchies to be learned. Mixture of Experts (MoE) offers a natural paradigm to address these challenges. Despite their potential, we demonstrate that naive MoEs are inadequate for the nuanced demands of hierarchical DSR, largely due to their gradient descent-based gating update mechanism which leads to slow updates and conflicted routing during training. To overcome this limitation, we introduce MixER: Mixture of Expert Reconstructors, a novel sparse top-1 MoE layer employing a custom gating update algorithm based on $K$-means and least squares. Extensive experiments validate MixER's capabilities, demonstrating efficient training and scalability to systems of up to ten parametric ordinary differential equations. However, our layer underperforms state-of-the-art meta-learners in high-data regimes, particularly when each expert is constrained to process only a fraction of a dataset composed of highly related data points. Further analysis with synthetic and neuroscientific time series suggests that the quality of the contextual representations generated by MixER is closely linked to the presence of hierarchical structure in the data.

cs.LG↗

arXiv2024

Stratify: Unifying Multi-Step Forecasting Strategies

A key aspect of temporal domains is the ability to make predictions multiple time steps into the future, a process known as multi-step forecasting (MSF). At the core of this process is selecting a forecasting strategy, however, with no existing frameworks to map out the space of strategies, practitioners are left with ad-hoc methods for strategy selection. In this work, we propose Stratify, a parameterised framework that addresses multi-step forecasting, unifying existing strategies and introducing novel, improved strategies. We evaluate Stratify on 18 benchmark datasets, five function classes, and short to long forecast horizons (10, 20, 40, 80). In over 84% of 1080 experiments, novel strategies in Stratify improved performance compared to all existing ones. Importantly, we find that no single strategy consistently outperforms others in all task settings, highlighting the need for practitioners explore the Stratify space to carefully search and select forecasting strategies based on task-specific requirements. Our results are the most comprehensive benchmarking of known and novel forecasting strategies. We make code available to reproduce our results.

cs.LG↗

arXiv2024

PFCM: Poisson flow consistency models for low-dose CT image denoising

X-ray computed tomography (CT) is widely used for medical diagnosis and treatment planning; however, concerns about ionizing radiation exposure drive efforts to optimize image quality at lower doses. This study introduces Poisson Flow Consistency Models (PFCM), a novel family of deep generative models that combines the robustness of PFGM++ with the efficient single-step sampling of consistency models. PFCM are derived by generalizing consistency distillation to PFGM++ through a change-of-variables and an updated noise distribution. As a distilled version of PFGM++, PFCM inherit the ability to trade off robustness for rigidity via the hyperparameter $D \in (0,\infty)$. A fact that we exploit to adapt this novel generative model for the task of low-dose CT image denoising, via a ``task-specific'' sampler that ``hijacks'' the generative process by replacing an intermediate state with the low-dose CT image. While this ``hijacking'' introduces a severe mismatch -- the noise characteristics of low-dose CT images are different from that of intermediate states in the Poisson flow process -- we show that the inherent robustness of PFCM at small $D$ effectively mitigates this issue. The resulting sampler achieves excellent performance in terms of LPIPS, SSIM, and PSNR on the Mayo low-dose CT dataset. By contrast, an analogous sampler based on standard consistency models is found to be significantly less robust under the same conditions, highlighting the importance of a tunable $D$ afforded by our novel framework. To highlight generalizability, we show effective denoising of clinical images from a prototype photon-counting system reconstructed using a sharper kernel and at a range of energy levels.

eess.IV↗

arXiv2024

Time-Series Classification for Dynamic Strategies in Multi-Step Forecasting

Multi-step forecasting (MSF) in time-series, the ability to make predictions multiple time steps into the future, is fundamental to almost all temporal domains. To make such forecasts, one must assume the recursive complexity of the temporal dynamics. Such assumptions are referred to as the forecasting strategy used to train a predictive model. Previous work shows that it is not clear which forecasting strategy is optimal a priori to evaluating on unseen data. Furthermore, current approaches to MSF use a single (fixed) forecasting strategy. In this paper, we characterise the instance-level variance of optimal forecasting strategies and propose Dynamic Strategies (DyStrat) for MSF. We experiment using 10 datasets from different scales, domains, and lengths of multi-step horizons. When using a random-forest-based classifier, DyStrat outperforms the best fixed strategy, which is not knowable a priori, 94% of the time, with an average reduction in mean-squared error of 11%. Our approach typically triples the top-1 accuracy compared to current approaches. Notably, we show DyStrat generalises well for any MSF task.

cs.LG↗

arXiv2024

A coronary artery phantom for task-based CT performance assessment and a comparative study of clinical CT, photon counting CT, and micro CT

While drastic improvements in CT technology have occurred in the past 25 years, spatial resolution is one area where progress has been limited until recently. New photon counting CT systems, are capable of much better spatial resolution than their (energy integrating) predecessors. These improvements have the potential to improve the evaluation obstructive coronary artery disease by enabling more accurate delineation between calcified plaque and coronary vessel lumen. A new set of vessel phantoms has been designed and manufactured for quantifying this improvement. Comparisons are made between an existing clinical CT system, a prototype photon counting system, with images from a micro CT system being used as the gold standard. Scans were made of the same objects on all three systems. The resulting images were registered and the luminal cross section areas were compared. Luminal cross-sections near calcified plaques were reduced due to blooming, but this effect was much less pronounced in images from the prototype photon counting system as compared to the images from the clinical CT system.

physics.med-ph↗

arXiv2021

AstronomicAL: An interactive dashboard for visualisation, integration and classification of data using Active Learning

AstronomicAL is a human-in-the-loop interactive labelling and training dashboard that allows users to create reliable datasets and robust classifiers using active learning. This technique prioritises data that offer high information gain, leading to improved performance using substantially less data. The system allows users to visualise and integrate data from different sources and deal with incorrect or missing labels and imbalanced class sizes. AstronomicAL enables experts to visualise domain-specific plots and key information relating both to broader context and details of a point of interest drawn from a variety of data sources, ensuring reliable labels. In addition, AstronomicAL provides functionality to explore all aspects of the training process, including custom models and query strategies. This makes the software a tool for experimenting with both domain-specific classifications and more general-purpose machine learning strategies. We illustrate using the system with an astronomical dataset due to the field's immediate need; however, AstronomicAL has been designed for datasets from any discipline. Finally, by exporting a simple configuration file, entire layouts, models, and assigned labels can be shared with the community. This allows for complete transparency and ensures that the process of reproducing results is effortless

astro-ph.IM↗