Jul 14, 2024 · Abstract:Existing methods for graph out-of-distribution (OOD) generalization primarily rely on empirical studies on synthetic datasets.

Mar 4, 2024 · In this paper, we propose a novel approach to generate hierarchical semantic environments for each graph.

Sep 21, 2024 · Existing methods for graph out-of-distribution (OOD) generalization primarily rely on empirical studies on synthetic datasets.

Out-of-distribution (OOD) generalization in the graph domain is challenging due to complex distribution shifts and a lack of environmental contexts.

Jul 15, 2024 · This paper introduces several novel techniques to improve the out-of-distribution generalization performance of graph neural networks on real-world data.

Mar 26, 2024 · Extensive experiments on synthetic and real-world OOD benchmarks demonstrate our method's effectiveness in improving OOD generalization. Our ...

GALA consistently improves the OOD generalization performance under various real-world graph distribution shifts on a number of realistic graph benchmarks: Page ...

Sep 23, 2023 · This paper presents a novel method to improve out-of-distribution (OOD) generalization for node-level tasks on graph data.

Following this strategy, to enhance the OOD generalization ability of models, we use the distribution shift information provided by the difference of training ...

We introduce the concept of co-adversarial perturbation optimization, which considers both model weights and node features, and we design an alternating ...