arXiv subjects

Zhongyuan Zhang

Publications and source records attributed to Zhongyuan Zhang.

7 recordsLinked to original sources

arXiv2026

Learning Multi-Relational Graph Representations for DNA Methylation-Based Biological Age Estimation

Aging clocks aim to estimate biological age, a measure of physiological state distinct from chronological age, from observable biomarkers, and are widely used for health assessment and disease analysis. DNA methylation is a particularly informative biomarker due to its stability and strong association with aging, and recent learning-based approaches have improved predictive performance. However, most existing methods treat CpG sites as independent features, overlooking the complex and heterogeneous biological relationships among them. We propose RelAge-GNN, a multi-relational graph neural network framework for DNA methylation-based age prediction. Our method constructs three complementary graphs capturing co-methylation patterns, genomic co-localization, and gene-level associations among CpG sites. Each graph is modeled by an independent GNN branch, and a learnable gating mechanism adaptively fuses the resulting representations. Experiments on large-scale datasets show that RelAge-GNN achieves competitive accuracy and stronger correlation with chronological age compared to state-of-the-art methods. Moreover, the model exhibits improved sensitivity in detecting age acceleration across diverse disease cohorts, highlighting its potential utility for disease characterization. Finally, through post hoc interpretability analyses, we quantify the contributions of different relational structures and CpG sites, providing biologically meaningful insights and suggesting potential directions for aging-related research. Our code is available at: https://anonymous.4open.science/r/RelAge-GNN-F1E3/.

cs.LG↗

arXiv2026

A nearly pristine star from the Large Magellanic Cloud

The first stars formed out of pristine gas, causing them to be so massive that none are expected to have survived until today. If their direct descendants were sufficiently low-mass stars, such stars could exist today and would be recognizable by having the lowest metallicities (abundance of elements heavier than helium). We present the independent identification and detailed chemical analysis of the star SDSS J0715-7334, finding ultra-low elemental abundances of both iron and carbon ([Fe/H] = -4.3, [C/Fe] < -0.2) and total metallicity Z < 7.8 x 10^{-7} (log Z/Zsun < -4.3). The star's orbit indicates that it originates from the halo of the Large Magellanic Cloud. Its heavy element abundance pattern can be explained by a primordial supernova with an initial mass of 30 solar masses. This star is over ten times more chemically pristine than the most extreme high-redshift galaxies currently found by the James Webb Space Telescope. It is sufficiently metal-poor that current models of low-mass star formation require dust cooling to explain its existence.

astro-ph.SR↗

arXiv2026

Prototype-Enhanced Multi-View Learning for Thyroid Nodule Ultrasound Classification

Thyroid nodule classification using ultrasound imaging is essential for early diagnosis and clinical decision-making; however, despite promising performance on in-distribution data, existing deep learning methods often exhibit limited robustness and generalisation when deployed across different ultrasound devices or clinical environments. This limitation is mainly attributed to the pronounced heterogeneity of thyroid ultrasound images, which can lead models to capture spurious correlations rather than reliable diagnostic cues. To address this challenge, we propose PEMV-thyroid, a Prototype-Enhanced Multi-View learning framework that accounts for data heterogeneity by learning complementary representations from multiple feature perspectives and refining decision boundaries through a prototype-based correction mechanism with mixed prototype information. By integrating multi-view representations with prototype-level guidance, the proposed approach enables more stable representation learning under heterogeneous imaging conditions. Extensive experiments on multiple thyroid ultrasound datasets demonstrate that PEMV-thyroid consistently outperforms state-of-the-art methods, particularly in cross-device and cross-domain evaluation scenarios, leading to improved diagnostic accuracy and generalisation performance in real-world clinical settings. The source code is available at https://github.com/chenyangmeii/Prototype-Enhanced-Multi-View-Learning.

cs.CV↗

arXiv2026

Learning Hierarchical Orthogonal Prototypes for Generalized Few-Shot 3D Point Cloud Segmentation

Generalized few-shot 3D point cloud segmentation aims to adapt to novel classes from only a few annotations while maintaining strong performance on base classes, but this remains challenging due to the inherent stability-plasticity trade-off: adapting to novel classes can interfere with shared representations and cause base-class forgetting. We present HOP3D, a unified framework that learns hierarchical orthogonal prototypes with an entropy-based few-shot regularizer to enable robust novel-class adaptation without degrading base-class performance. HOP3D introduces hierarchical orthogonalization that decouples base and novel learning at both the gradient and representation levels, effectively mitigating base-novel interference. To further enhance adaptation under sparse supervision, we incorporate an entropy-based regularizer that leverages predictive uncertainty to refine prototype learning and promote balanced predictions. Extensive experiments on ScanNet200 and ScanNet++ demonstrate that HOP3D consistently outperforms state-of-the-art baselines under both 1-shot and 5-shot settings. The code is available at https://fdueblab-hop3d.github.io/.

cs.CV↗

arXiv2026

Quantifying Element Importance for Mass Recovery from Population III Supernova Yield Fits

Massive Population III stars are currently not observed, but their initial mass function (IMF) can be inferred through stellar archaeology: fitting core-collapse supernova yield models to elemental abundances of low-mass, long-lived metal-poor stars. While prior work demonstrates that yield fitting can recover progenitor properties, it remains unclear which measured elements most control mass recovery quality and what level of IMF precision is achievable for a measured element set. We perform a systematic study of element importance for progenitor mass recovery. Using the Heger & Woosley (2010) yield grid, we generate mock observations, fit the initial mass, and evaluate the typical performance on the fractional mass recovery. Add/remove-one-element experiments and comparisons among different baseline element sets are used to rank elements by importance. We find that the most important elements for accurate mass recovery are C, N, Na, and K, with O, Al, Co, and Ni consistently improving performance when available. Overall, with currently measurable elements from high-resolution spectroscopy, stellar archaeology can deliver practical Population III IMF constraints assuming the core-collapse supernova yield models provide a good representation of stellar evolution in the early universe.

astro-ph.SR↗

arXiv2025

Hybrid Quantum Physics-informed Neural Network: Towards Efficient Learning of High-speed Flows

This study benchmarks hybrid quantum physics-informed neural network (HQPINN) to model high-speed flows, compared against classical physics-informed neural networks (PINNs) and fully quantum neural networks (QNNs). The HQPINN architecture integrates a parameterized quantum circuit (PQC) with a classical neural network in parallel, trained via a physics-informed loss. Across harmonic, non-harmonic, and transonic benchmarks, HQPINNs demonstrate balanced performance, offering competitive accuracy and stability with reduced parameter cost. Quantum PINNs are highly efficient for harmonic problems achieving the lowest loss with minimal parameters due to their Fourier structure, but struggle to generalize in non-harmonic settings involving shocks and discontinuities. HQPINNs mitigate such artifacts, and with sufficient parameterization, can match the performance of classical models in more complex regimes. Although constrained by current quantum emulation costs and scalability, HQPINNs show promise as general-purpose solvers, offering parameter efficiency with robust fallback behavior, particularly suited for problems where the nature of the solution is not known a-priori.

physics.comp-ph↗

arXiv2022

Noise and Edge Based Dual Branch Image Manipulation Detection

Unlike ordinary computer vision tasks that focus more on the semantic content of images, the image manipulation detection task pays more attention to the subtle information of image manipulation. In this paper, the noise image extracted by the improved constrained convolution is used as the input of the model instead of the original image to obtain more subtle traces of manipulation. Meanwhile, the dual-branch network, consisting of a high-resolution branch and a context branch, is used to capture the traces of artifacts as much as possible. In general, most manipulation leaves manipulation artifacts on the manipulation edge. A specially designed manipulation edge detection module is constructed based on the dual-branch network to identify these artifacts better. The correlation between pixels in an image is closely related to their distance. The farther the two pixels are, the weaker the correlation. We add a distance factor to the self-attention module to better describe the correlation between pixels. Experimental results on four publicly available image manipulation datasets demonstrate the effectiveness of our model.

cs.CV↗