Graph neural networks trained on null signal for angle reconstruction in x-ray polarimetry
Open Access
- 1 June 2026
- journal article
- Published by IOP Publishing in Machine Learning: Science and Technology
- Vol. 7 (3) , 035063
Abstract
Scientific real detectors often produce sparse, irregular data defined on non-Euclidean domains, where conventional convolutional neural networks (CNNs) impose geometric biases that distort physical observables. In x-ray polarimetry, such distortions can mimic polarization signals at the percent level, critically limiting measurement fidelity. We consider a framework based on graph neural networks (GNNs) for reconstructing photoelectron emission angles from Gas Pixel Detectors, operating directly on their native sparse hexagonal topology. The network is trained solely on unpolarized simulated data and integrates rotational data augmentation, ensemble averaging, and modulation-aware model selection to reduce spurious angular modulation. Despite never encountering polarized examples, the model is able to recover the overall structure of polarized signals. Compared with the classical method of moments, an analytical approach to track angle reconstruction, the proposed GNN achieves lower per-track error but yields worse performance for polarimetry, revealing a mismatch between local reconstruction accuracy and global polarimetric performance. Compared to CNN-based approaches, the framework shows competitive performance on both unpolarized and polarized reconstructions, despite being trained in a more challenging power-law energy distribution. These results position the method as a diagnostic tool to highlight the limitations and trade-offs of learning-based approaches for x-ray-polarimetry signal reconstruction.Keywords
All Related Versions
Funding Information
- Italian Space Agency
This publication has 17 references indexed in Scilit:
- Optimization of deep learning method on track reconstruction for X-ray polarimetry with gas pixel detectorsExperimental Astronomy, 2025
- Research on the X-ray polarization deconstruction method based on hexagonal convolutional neural networkNuclear Science and Techniques, 2025
- Graph neural networks in vision-language image understanding: a surveyThe Visual Computer, 2024
- ixpeobssim: A simulation and analysis framework for the imaging X-ray polarimetry explorerSoftwareX, 2022
- Design, construction, and test of the Gas Pixel Detectors for the IXPE missionAstroparticle Physics, 2021
- Deep ensemble analysis for Imaging X-ray PolarimetryNuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2021
- A convolutional neural network approach for reconstructing polarization information of photoelectric X-ray polarimetersNuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2019
- A sealed Gas Pixel Detector for X-ray astronomyNuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2007
- Direct reading of charge multipliers with a self-triggering CMOS analog chip with 105k pixels at 50μm pitchNuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2006
- Astronomical X-ray polarimetry based on photoelectric effect with microgap detectorsNuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2001