Ciyuan Peng - Service and Talks

🎨Service

Invited Reviewer

International Joint Conference on Artificial Intelligence (IJCAI), 2025
The Web Conference (WWW), 2026
ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2026
Conference on Neural Information Processing Systems (NeurIPS), 2025-2026
International Conference on Learning Representations (ICLR), 2026-2027
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
ACM Transactions on Knowledge Discovery from Data (TKDD)
IEEE Transactions on Knowledge and Data Engineering (TKDE)
IEEE Transactions on Parallel and Distributed Systems (TPDS)
Information Science
IEEE Transactions on Neural Networks and Learning Systems (TNNLS)

Program Committee Member

The First Workshop on Graph Learning (GL 2022).
International workshop on knowledge graphs (KG2022)

🎙️Talks

I present our paper “Adaptive Brain Network Augmentation based on Group-aware Graph Learning” at ICLR 2024.

Abstract:

Brain network analysis significantly improves artificial intelligence techniques in the realm of digital health. Most existing methods uniformly construct brain networks for different groups (e.g., male and female groups, healthy and sick people groups), facing the interference of group-irrelevant noises and failing to capture group-specific features to enhance brain networks. To address this issue, this paper proposes an adaptive brain network augmentation method based on group-aware graph learning. We construct group-aware brain networks, which can adapt to distinct groups, reducing the interference of noises, and improving model robustness across various tasks and subject groups.

I presented our paper “Stage-aware Brain Graph Learning for Alzheimer′ s Disease” at CAI 2024.

Abstract:

Current machine learning-based Alzheimer’s disease (AD) diagnosis methods fail to explore the distinctive brain patterns across different AD stages, lacking the ability to trace the trajectory of AD progression. This limitation can lead to an oversight of the pathological mechanisms of AD and suboptimal performance in AD diagnosis. To overcome this challenge, this paper proposes a novel stage-aware brain graph learning model. Particularly, we analyze the different brain patterns of each AD stage in terms of stage-specific brain graphs. We design a Stage Feature-enhanced Graph Contrastive Learning method, named SF-GCL, utilizing specific features within each AD stage to perform graph augmentation, thereby effectively capturing differences between stages. Significantly, this study unveils the specific brain patterns corresponding to each AD stage, showing great potential in tracing the trajectory of brain degeneration. Experimental results on a real-world dataset demonstrate the superiority of our model.

I gave a talk on my PhD research “Multimodal Brain Graph Representation Learning” at AJCAI 2024.

Abstract:

We aim to develop a series of trustworthy models for multimodal brain graph representation learning. The overarching goal is to improve intelligent brain disease diagnosis, treatment, and brain activity analysis, while advancing the understanding of underlying brain mechanisms. In this thesis, trustworthiness is operationalized through three dimensions: biological plausibility, high multimodal representation quality, and technical reliability. Biological plausibility refers to the preservation of established brain properties. High multimodal representation quality refers to the effective modeling of meaningful relationships between different brain modalities. Technical reliability refers to reliable representation learning by mitigating essential information loss during multimodal fusion.

I presented our paper “Biologically Plausible Brain Graph Transformer” at ICLR 2025.

Abstract:

State-of-the-art brain graph analysis methods fail to fully encode the small-world architecture of brain graphs (accompanied by the presence of hubs and functional modules), and therefore lack biological plausibility to some extent. This limitation hinders their ability to accurately represent the brain’s structural and functional properties, thereby restricting the effectiveness of machine learning models in tasks such as brain disorder detection. In this work, we propose a novel Biologically Plausible Brain Graph Transformer (BioBGT) that encodes the small-world architecture inherent in brain graphs. Specifically, we present a network entanglement-based node importance encoding technique that captures the structural importance of nodes in global information propagation during brain graph communication, highlighting the biological properties of the brain structure. Furthermore, we introduce a functional module-aware self-attention to preserve the functional segregation and integration characteristics of brain graphs in the learned representations. Experimental results on three benchmark datasets demonstrate that BioBGT outperforms state-of-the-art models, enhancing biologically plausible brain graph representations for various brain graph analytical tasks.