arXiv:2510. 09416v4 Announce Type: replace Abstract: Learning on temporal graphs has become a central topic in graph representation learning, with numerous benchmarks indicating the strong performance of state-of-the-art models.
arXiv:2608. 07158v1 Announce Type: new Abstract: Temporal graph learning has become essential for analyzing real-world systems whose interactions continuously evolve over time, including financial transaction networks, communication systems, and online social platforms.
The article surveys Dynamic Heterogeneous Graph Representation Learning (DHGRL), a field that tackles the challenges of modeling evolving, multi‑type networks. It introduces a unified definition covering both discrete‑time and continuous‑time DHGs, and proposes an algorithm‑centric taxonomy that groups methods into embedding‑based, GNN‑based, and Transformer‑based approaches, highlighting their biases toward temporal granularity. The survey also reviews key applications, datasets, benchmarks, and outlines future research directions.
arXiv:2605.06462v2 Announce Type: replace Abstract: Progress in graph learning is hindered by benchmark practices that conflate the contributions of node features and graph structure, making it hard...
The paper explores how information propagates in Deep Graph Networks (DGNs) for both static and dynamic graphs, treating DGNs as dynamical systems. It presents new architectures that better preserve long‑term node dependencies and learn complex spatio‑temporal patterns from irregular, sparsely sampled dynamic graphs. The work combines theoretical analysis with empirical results to demonstrate the effectiveness of these designs.
arXiv:2607. 07716v1 Announce Type: cross Abstract: Temporal graphs are ubiquitous in real-world applications and Temporal Graph Networks (TGNs) have achieved superior predictive accuracy.
TTGBench is a new benchmark for temporal graph learning that evaluates both structural evolution and semantic drift in text‑attributed graphs. It includes six real‑world, text‑rich datasets with dual volatility and supports multi‑class and multi‑label temporal node classification, addressing gaps left by existing benchmarks. A comprehensive evaluation of 17 state‑of‑the‑art methods shows a clear divide: TGNNs excel at structural prediction but struggle with semantic tracking, while LLM‑based models perform better on semantic tasks but lag in structural prediction.
arXiv:2405. 19062v2 Announce Type: replace-cross Abstract: Continuous-Time Dynamic Graphs (CTDGs) enable fine-grained modeling of evolving relational systems.
The paper introduces the Graph Dynamics Model (GDM), a world model that learns stochastic latent dynamics over evolving graph topologies. GDM employs a sparse recurrent adjacency matrix for topology updates and a recurrent state‑space architecture for stochastic transitions, enabling it to handle partially observable, stochastic environments. The authors also propose the Graph Distribution Distance (GDD) metric, using maximum mean discrepancy with a graph kernel, to compare predicted and true joint graph state distributions, and demonstrate GDM’s superior performance and zero‑shot generalisation on large graphs.
arXiv:2602. 14239v3 Announce Type: replace-cross Abstract: Predicting links in sparse, continuously evolving networks is a central challenge in network science.
arXiv:2608. 03696v1 Announce Type: new Abstract: This work focuses on the problem of learning on temporal graphs, with particular emphasis on the task of clustering: obtaining coarse-grained representations by aggregating information from nodes, edges, and temporal dynamics - a task related to pooling in machine learning on graphs, or community detection in network science.
CacheDyG introduces a cache‑refine framework that decouples temporal propagation from parameter updates in dynamic graph neural networks. By storing graph‑aware node‑time representations in non‑trainable buffers and updating only a lightweight refiner, residual gate, and link predictor during training, it reduces repeated recomputation of historical structures. Experiments on five benchmarks show that CacheDyG uses fewer trainable parameters, runs faster, and achieves competitive or better predictive performance compared to existing baselines.
arXiv:2506. 00431v2 Announce Type: replace Abstract: Due to the proficiency of self-attention mechanisms (SAMs) in capturing dependencies in sequence modeling, several existing dynamic graph neural networks (DGNNs) utilize Transformer architectures with various encoding designs to capture sequential evolutions of dynamic graphs.