Deep learning relies on learning from extensive data to generate prediction results. This
approach may inadvertently capture spurious correlations within the data, leading to models …

Abstract Graph Neural Networks (GNNs) have been proven to easily overfit spurious
subgraphs in the available data, which reduces their trustworthiness in high-stakes real …

Actual graph data typically encompasses multiple types and evolves over time, thus they are
commonly represented as dynamic heterogeneous graphs. In recent years, dynamic …

To break the scalability constraint of GNNs, distilling graph knowledge to student MLPs
provides a promising strategy for inferring on large-scale graph tasks. Due to the fact that …

Graph Neural Networks (GNNs) have emerged as powerful representation learning tools for
capturing complex dependencies within diverse graph-structured data. Despite their …

Signal processing is crucial for satisfying the high data rate requirements of future sixth-
generation (6G) wireless networks. However, the rapid growth of wireless networks has …

To address the limitations of existing graph contrastive learning methods, which fail to
adaptively integrate feature and topological information and struggle to efficiently capture …

In high-energy physics, particle jet tagging plays a pivotal role in distinguishing quark from
gluon jets using data from collider experiments. While graph-based deep learning methods …

Molecular representation learning is widely used in the field of drug discovery, due to its
ability to accurately capture the complex features of compounds in high-dimensional space …