Good: A graph out-of-distribution benchmark

S Gui, X Li, L Wang, S Ji - Advances in Neural Information …, 2022 - proceedings.neurips.cc
Abstract Out-of-distribution (OOD) learning deals with scenarios in which training and test
data follow different distributions. Although general OOD problems have been intensively …

Dynamic graph neural networks under spatio-temporal distribution shift

Z Zhang, X Wang, Z Zhang, H Li… - Advances in neural …, 2022 - proceedings.neurips.cc
Dynamic graph neural networks (DyGNNs) have demonstrated powerful predictive abilities
by exploiting graph structural and temporal dynamics. However, the existing DyGNNs fail to …

Unleashing the power of graph data augmentation on covariate distribution shift

Y Sui, Q Wu, J Wu, Q Cui, L Li, J Zhou… - Advances in Neural …, 2023 - proceedings.neurips.cc
The issue of distribution shifts is emerging as a critical concern in graph representation
learning. From the perspective of invariant learning and stable learning, a recently well …

Environment-aware dynamic graph learning for out-of-distribution generalization

H Yuan, Q Sun, X Fu, Z Zhang, C Ji… - Advances in Neural …, 2023 - proceedings.neurips.cc
Dynamic graph neural networks (DGNNs) are increasingly pervasive in exploiting spatio-
temporal patterns on dynamic graphs. However, existing works fail to generalize under …

Spectral invariant learning for dynamic graphs under distribution shifts

Z Zhang, X Wang, Z Zhang, Z Qin… - Advances in …, 2023 - proceedings.neurips.cc
Dynamic graph neural networks (DyGNNs) currently struggle with handling distribution shifts
that are inherent in dynamic graphs. Existing work on DyGNNs with out-of-distribution …

A Simple Data Augmentation for Graph Classification: A Perspective of Equivariance and Invariance

Y Sui, S Wang, J Sun, Z Liu, Q Cui, L Li, J Zhou… - ACM Transactions on …, 2024 - dl.acm.org
In graph classification, the out-of-distribution (OOD) issue is attracting great attention. To
address this issue, a prevailing idea is to learn stable features, on the assumption that they …

Graph domain adaptation with dual-branch encoder and two-level alignment for whole slide image-based survival prediction

Y Shou, P Yan, X Yuan, X Cao, Q Zhao… - arxiv preprint arxiv …, 2024 - arxiv.org
In recent years, histopathological whole slide image (WSI)-based survival analysis has
attracted much attention in medical image analysis. In practice, WSIs usually come from …

Out-of-distribution generalized dynamic graph neural network with disentangled intervention and invariance promotion

Z Zhang, X Wang, Z Zhang, H Li, W Zhu - arxiv preprint arxiv:2311.14255, 2023 - arxiv.org
Dynamic graph neural networks (DyGNNs) have demonstrated powerful predictive abilities
by exploiting graph structural and temporal dynamics. However, the existing DyGNNs fail to …

Wasserstein distance regularized graph neural networks

Y Shi, L Zheng, P Quan, L Niu - Information Sciences, 2024 - Elsevier
Distribution shift widely exists in graph representation learning and often reduces model
performance. This work investigates how to improve the performance of a graph neural …

Mario: Model agnostic recipe for improving ood generalization of graph contrastive learning

Y Zhu, H Shi, Z Zhang, S Tang - … of the ACM Web Conference 2024, 2024 - dl.acm.org
In this work, we investigate the problem of out-of-distribution (OOD) generalization for
unsupervised learning methods on graph data. To improve the robustness against such …