arXiv subjects

Andrew Hope

Publications and source records attributed to Andrew Hope.

3 recordsLinked to original sources

arXiv2026

A Hybrid R-Theta Fiber Positioning Robot Using a Piezoelectric Bender for Radial Motion

Future massively multiplexed spectroscopic surveys require focal planes with increasingly dense arrays of robotic fiber positioners. The DESI (Dark Energy Spectroscopic Instrument) fiber positioners, based on a two-axis Theta-Phi motorized architecture, have demonstrated excellent reliability and performance during operations and provide the benchmark for evaluating new concepts. However, proposed next-generation instruments such as Spec-S5 require a substantially smaller positioner pitch than DESI, making direct scaling of the existing design challenging. Here we describe a hybrid R-Theta fiber positioning concept in which the rotation about the central axis is provided by a DESI-like Theta motor, while the radial displacement is produced by a piezoelectric bimorph bender coupled to a lightweight carbon-fiber spine. This design builds on the heritage of the the proven central rotation mechanism while replacing the Phi arm with a compact piezoelectric actuator. We discuss the trade-offs between motorized positioners, piezoelectric slip-stick tilting spines, tubular piezo actuators, and direct piezo-bender actuation. A prototype based on a DESI positioner was fabricated and tested using a commercial bimorph bender and an 8-inch carbon-fiber extension. With closed-loop control using feedback from a fiber-view camera the fiber was brought on target after two corrections with residual stability of approximately 2 microns RMS. These results indicate that direct piezo-bender actuation is a promising approach for compact fiber positioning, provided that closed-loop operation is addressed.

astro-ph.IM↗

arXiv2025

CAST: Time-Varying Treatment Effects with Application to Chemotherapy and Radiotherapy on Head and Neck Squamous Cell Carcinoma

Causal machine learning (CML) enables individualized estimation of treatment effects, offering critical advantages over traditional correlation-based methods. However, existing approaches for medical survival data with censoring such as causal survival forests estimate effects at fixed time points, limiting their ability to capture dynamic changes over time. We introduce Causal Analysis for Survival Trajectories (CAST), a novel framework that models treatment effects as continuous functions of time following treatment. By combining parametric and non-parametric methods, CAST overcomes the limitations of discrete time-point analysis to estimate continuous effect trajectories. Using the RADCURE dataset [1] of 2,651 patients with head and neck squamous cell carcinoma (HNSCC) as a clinically relevant example, CAST models how chemotherapy and radiotherapy effects evolve over time at the population and individual levels. By capturing the temporal dynamics of treatment response, CAST reveals how treatment effects rise, peak, and decline over the follow-up period, helping clinicians determine when and for whom treatment benefits are maximized. This framework advances the application of CML to personalized care in HNSCC and other life-threatening medical conditions. Source code/data available at: https://github.com/CAST-FW/HNSCC

cs.LG↗

arXiv2021

A Machine Learning Challenge for Prognostic Modelling in Head and Neck Cancer Using Multi-modal Data

Accurate prognosis for an individual patient is a key component of precision oncology. Recent advances in machine learning have enabled the development of models using a wider range of data, including imaging. Radiomics aims to extract quantitative predictive and prognostic biomarkers from routine medical imaging, but evidence for computed tomography radiomics for prognosis remains inconclusive. We have conducted an institutional machine learning challenge to develop an accurate model for overall survival prediction in head and neck cancer using clinical data etxracted from electronic medical records and pre-treatment radiological images, as well as to evaluate the true added benefit of radiomics for head and neck cancer prognosis. Using a large, retrospective dataset of 2,552 patients and a rigorous evaluation framework, we compared 12 different submissions using imaging and clinical data, separately or in combination. The winning approach used non-linear, multitask learning on clinical data and tumour volume, achieving high prognostic accuracy for 2-year and lifetime survival prediction and outperforming models relying on clinical data only, engineered radiomics and deep learning. Combining all submissions in an ensemble model resulted in improved accuracy, with the highest gain from a image-based deep learning model. Our results show the potential of machine learning and simple, informative prognostic factors in combination with large datasets as a tool to guide personalized cancer care.

cs.LG↗