State of the Field
Recent advancements in graph learning are increasingly focused on enhancing the representation and utility of multimodal and dynamic networks. Researchers are exploring novel frameworks that integrate diverse data types, such as the development of Clifford neural paradigms for multimodal-attributed graphs, which improve modality alignment and fusion. Additionally, counterfactual data augmentation techniques are being employed to bolster the predictive capabilities of dynamic networks, allowing models to adapt to evolving structures without extensive architectural changes. The exploration of higher-order relational structures through topological deep learning is also gaining traction, with new models designed to efficiently propagate information in complex networks. Furthermore, the concept of a graph substrate is emerging, emphasizing the need for persistent structural representations that can be shared across various tasks and modalities. Collectively, these efforts aim to address practical challenges in scalability, robustness, and generalizability, positioning graph learning as a versatile tool for solving complex real-world problems across diverse domains.
Papers
1–7 of 7LION: A Clifford Neural Paradigm for Multimodal-Attributed Graph Learning
Recently, the rapid advancement of multimodal domains has driven a data-centric paradigm shift in graph ML, transitioning from text-attributed to multimodal-attributed graphs. This advancement signifi...
CCMamba: Selective State-Space Models for Higher-Order Graph Learning on Combinatorial Complexes
Topological deep learning has emerged for modeling higher-order relational structures beyond pairwise interactions that standard graph neural networks fail to capture. Although combinatorial complexes...
CoDCL: Counterfactual Data Augmentation Contrastive Learning for Continuous-Time Dynamic Network Link Prediction
The rapid growth and continuous structural evolution of dynamic networks make effective predictions increasingly challenging. To enable prediction models to adapt to complex temporal environments, the...
DyGnROLE: Modeling Asymmetry in Dynamic Graphs with Node-Role-Oriented Latent Encoding
Real-world dynamic graphs are often directed, with source and destination nodes exhibiting asymmetrical behavioral patterns and temporal dynamics. However, existing dynamic graph architectures largely...
Simple Network Graph Comparative Learning
The effectiveness of contrastive learning methods has been widely recognized in the field of graph learning, especially in contexts where graph data often lack labels or are difficult to label. Howeve...
Graph is a Substrate Across Data Modalities
Graphs provide a natural representation of relational structure that arises across diverse domains. Despite this ubiquity, graph structure is typically learned in a modality- and task-isolated manner,...
RiemannGL: Riemannian Geometry Changes Graph Deep Learning
Graphs are ubiquitous, and learning on graphs has become a cornerstone in artificial intelligence and data mining communities. Unlike pixel grids in images or sequential structures in language, graphs...