Aug 2026· Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2· pp. 5845-5855· 0 citations· 21 references
Abstract
Spatio-temporal graph forecasting is crucial for numerous real-world applications but often suffers from severe data scarcity in target domains, necessitating knowledge transfer from data-rich sources. In this work, we push this to the extreme by studying zero-shot spatio-temporal graph learning, where only the target graph topology is accessible and no temporal observations from the target domain are available at all. This strict setting renders state-of-the-art node-adaptive spatio-temporal graph neural networks ineffective, as their node-specific parameters cannot be inferred without target data. To bridge this gap, we propose \sysname, an Orbit-Adaptive Graph Neural Network framework. Our key insight is that a node's temporal behavior is not only determined by its own historical series but is also informed by its structural role within the graph, which we formally characterize via automorphism orbits—an equivalence relation capturing symmetric positions in the graph. By sharing adaptation mechanisms across nodes belonging to the same orbit, \sysname learns domain-invariant, structure-aware representations that enable prediction on completely unseen graphs. Extensive experiments on six datasets across traffic, air quality, and pandemic domains demonstrate that \sysname outperforms existing transfer and naive zero-shot baselines, establishing a new paradigm for structure-driven zero-shot forecasting.
DynaSTar is proposed, a Dynamic Spatio-Temporal Graph Invariant Learning model designed for reliable out-of-time (OOT) traffic prediction under evolving topologies, which employs a dynamic probabilistic graph structure, which is continuously refined through momentum-based updates and differentiable sparse sampling to model evolving inter-node dependencies.
Xinyan Hao, Huaiyu Wan, S. Guo et al.· 0 citations
Spatio-temporal graphs are widely used in modeling complex dynamic processes such as temporal forecasting, molecular dynamics, and healthcare monitoring. Recently, stringent privacy regulations such as GDPR and CCPA have introduced significant new challenges for existing spatio-temporal graph models, requiring complete unlearning of unauthorized data. Since each node in a spatio-temporal graph diffuses information globally across both spatial and temporal dimensions, existing unlearning methods primarily designed for static graphs and localized data removal cannot efficiently erase a single node without incurring costs nearly equivalent to full model retraining. To address this, we propose CallosumNet, a spatio-temporal graph unlearning framework biologically inspired by the corpus callosum structure. CallosumNet makes two key technical contributions: (1) it reconstructs subgraphs using biologically-inspired virtual edges; and (2) it restores interlinked spatio-temporal dependencies among subgraphs via a lightweight meta-graph integration layer. Empirical results on four diverse real-world datasets show that CallosumNet achieves complete unlearning while maintaining accuracy very close to the gold model. The code is publicly available at https://github.com/wenlu-lab/STGraphUnlearning.
Qi-Ming Guo, Wenbo Sun, Chen Pan et al.· 0 citations
This paper proposes FAST-GNN, a graph neural network that integrates frequency-aware and spatio-temporal adaptive mechanisms to address the challenges of modeling complex spatio-temporal dependencies in multivariate time series forecasting. Traditional GNNs often struggle with dynamic graph construction, frequency-domain feature extraction, and noise robustness. Our approach innovatively introduces a multi-band driven dynamic graph construction mechanism to finely capture multi-scale correlations, employs a dual-granularity adaptive graph filter to distinguish and integrate local and global dependencies, and adopts a hierarchical graph reading strategy to enhance robustness against noise and anomalies. Extensive experiments on real-world datasets, including Climate, Electricity, Weather, and PEMS series, demonstrate that FAST-GNN significantly outperforms state-of-the-art baselines in both long- and short-term forecasting tasks, particularly showing stable and superior performance in challenging scenarios such as high-noise environments and long-term horizons.
Shuming Zhang· International Conference on...· 0 citations
It is shown that high-frequency fluctuations at a single station are dominated by instrumental and stochastic noise, making temporal filtering via downsampling essential, and utilizing a learned adjacency matrix as an adaptable structural regularization provides a highly stable and effective framework for short-term geomagnetic field forecasting.
Bryan Tipán, E. López, W. Carvajal et al.· Frontiers in Astronomy and S...· 0 citations
A deeply fused GraphSAGE-GRU cell that embeds independent inductive GraphSAGE(SAmple and aggreGatE) encoders directly into each GRU gate, enabling simultaneous spatio-temporal feature extraction at every time step while remaining topology-agnostic.
Xuran Chen· Poster Volume 0008 The 2026...· 0 citations
Probabilistic spatio-temporal forecasting is a pivotal challenge in the data mining community. Recently, diffusion models have emerged as a powerful paradigm in this field, acting as conditional generative models. Beyond standard diffusion models, a significant research trend involves tailoring the diffusion model to incorporate domain-specific inductive biases. However, for complex spatio-temporal data, the high-dimensional coupling between spatial and temporal domains makes the design of such models exceptionally difficult. In the present work, we propose a novel framework (GODM) that adopts orthogonalization to decouple intricate temporal dynamics while employing spatial convolutions to simulate information flow across space. By embedding the spatio-temporal structure directly into the forward process, the GODM introduces crucial inductive biases with negligible computational overhead. Extensive experiments on multiple benchmark datasets demonstrate that the GODM framework consistently outperforms state-of-the-art spatio-temporal diffusion models.
Zhixian Wang, Michel Ferreira Cardia Haddad, Jose Eduardo Medina Reyes et al.· Proceedings of the 32nd ACM...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
