Peng Zhang

Peng Zhang

Auto-generated profile, based on publicly available information

Peng Zhang

Cyberspace Institute of Advanced Technology, Guangzhou University, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 April 2023
Verified
Conference Paper
Open access
YLYang Liu
Yang Liu
CZChuan Zhou
Chuan Zhou
SPShirui Pan
Shirui Pan
Jia WUJia WU
Jia WU
ZLZhao Li
Zhao Li
PZPeng Zhang
Peng Zhang
Graph neural networks (GNNs) are powerful models to handle graph data and can achieve state-of-the-art in many critical tasks including node classification and link prediction. However, existing graph neural networks still face both challenges of over-smoothing and over-squashing based on previous literature. To this end, we propose a new Curvature-based topology-aware Dropout sampling technique named CurvDrop, in which we integrate the Discrete Ricci Curvature into graph neural networks to enable more expressive graph models. Also, this work can improve graph neural networks by quantifying connections in graphs and using structural information such as community structures in graphs. As a result, our method can tackle the both challenges of over-smoothing and over-squashing with theoretical justification. Also, numerous experiments on public datasets show the effectiveness and robustness of our proposed method. The code and data are released in https://github.com/liu-yang-maker/Curvature-based-Dropout.

26 March 2025 in Neural Networks
Verified
Journal Article
Closed access
YCYanan Cao
Yanan Cao
FSFengzhao Shi
Fengzhao Shi
QYQing Yu
Qing Yu
XLXixun Lin
Xixun Lin
CZChuan Zhou
Chuan Zhou
DYDawei Yin
Dawei Yin

30 April 2023
Verified
Conference Paper
Open access
YZYuchen Zhou
Yuchen Zhou
YCYanan Cao
Yanan Cao
Yongchao LiuYongchao Liu
Yongchao Liu
YSYanmin Shang
Yanmin Shang
PZPeng Zhang
Peng Zhang
WWWeiqiang Wang
Weiqiang Wang
Data augmentation has been widely studied as it can be used to improve the generalizability of graph representation learning models. However, existing works focus only on the data augmentation on homogeneous graphs. Data augmentation for heterogeneous graphs remains under-explored. Considering that heterogeneous graphs contain different types of nodes and links, ignoring the type information and directly applying the data augmentation methods of homogeneous graphs to heterogeneous graphs will lead to suboptimal results. In this paper, we propose a novel Multi-Aspect Heterogeneous Graph Augmentation framework named MAHGA. Specifically, MAHGA consists of two core augmentation strategies: structure-level augmentation and metapath-level augmentation. Structure-level augmentation pays attention to network schema aspect and designs a relation-aware conditional variational auto-encoder that can generate synthetic features of neighbors to augment the nodes and the node types with scarce links. Metapath-level augmentation concentrates on metapath aspect, which constructs metapath reachable graphs for different metapaths and estimates the graphons of them. By sampling and mixing up based on the graphons, MAHGA yields intra-metapath and inter-metapath augmentation. Finally, we conduct extensive experiments on multiple benchmarks to validate the effectiveness of MAHGA. Experimental results demonstrate that our method improves the performances across a set of heterogeneous graph learning models and datasets.

28 August 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
HYHong Yang
Hong Yang
RHRuiwen Huang
Ruiwen Huang
PZPeng Zhang
Peng Zhang
YZYao Zhang
Yao Zhang
YZYanchun Zhang
Yanchun Zhang
Multimodal learning has been extensively exploited for brain disease diagnosis. However, integrating structural and functional magnetic resonance imaging (sMRI and fMRI) for reliable brain disease diagnosis remains a significant challenge, because unreliable cross-modal information may exist during fusion learning which leads to large mismatch between feature representation and reconstruction learning of sMRI and fMRI data. In this paper, we propose a new Reliable Multimodal Fusion Network (RMFN for short) that jointly learns spatial brain representations, temporal dynamics, and reliable cross-modal fusion. First, RMFN encodes fMRI data using Rotary Positional Embedding-enhanced Transformers to capture temporal dependencies within fMRI data. Then, RMFN represents sMRI data as region-of-interest (ROI) graphs and uses Graph Attention Networks to learn spatial relationships among brain regions. Lastly, a reliable cross-modal learning function is developed to combine both sMRI and fMRI data through adaptive weighting based on feature representation and reconstruction learning. Experiments show the utility of the proposed model. Interpretability analysis show neurobiologically significant biomarkers which are consistent with established clinical findings.

26 September 2026 in Neural Networks
Verified
Journal Article
Closed access
GTGusheng Tian
Gusheng Tian
HYHong Yang
Hong Yang
PZPeng Zhang
Peng Zhang