Hierarchical Spatio-Temporal Graph Neural Networks for Pandemic Forecasting

Ma, Yihong; Gerard, Patrick D.; Tian, Yijun; Guo, Zhu Ming; Chawla, Nitesh V.

doi:10.1145/3511808.3557350

Cited by 12 publications

(5 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hierarchical Spatio-Temporal Data Mining. Existing hierarchical spatio-temporal data mining works implement hierarchy construction by clustering node representations [7,17,18,28,29,32]. Traffic-Transformer [32] and ST-HSL [17] capture global node dependencies using multi-head attention mechanisms and hypergraphs.…”

Section: Related Workmentioning

confidence: 99%

“…Traffic-Transformer [32] and ST-HSL [17] capture global node dependencies using multi-head attention mechanisms and hypergraphs. HGNN [29] and HiGCN [18] generate new hierarchies using unique trainable GAT or GCN models, while MC-STGCN [28] uses the Louvain algorithm [4] to detect potential communities.…”

Section: Related Workmentioning

confidence: 99%

“…Instead, our method efficiently solves the BCC to merge sensors with high intra-region dependency and generate the regional map with linear time complexity, which does not require updating or training and can be easily extended. Other works [18,29] construct unique graph neural networks for different hierarchies. In contrast, we incorporate a Meta GCN.…”

Section: Related Workmentioning

confidence: 99%

“…Nevertheless, this series of works treats all fine-grained sensors equally regardless of the macro dependency among the sensors and their high-order neighbors. Recently, efforts have been exerted to explore hierarchical sensor modeling [7,17,18,28,29,32]. HGCN [7] devises a hierarchical view based on the fine-grained sensors by simply clustering the node representations.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Rethinking Sensors Modeling: Hierarchical Information Enhanced Traffic Forecasting

Ma,

Zhang,

Zhao

et al. 2023

Proceedings of the 32nd ACM International Conference on Information and Knowledge Management

View full text Add to dashboard Cite

With the acceleration of urbanization, traffic forecasting has become an essential role in smart city construction. In the context of spatio-temporal prediction, the key lies in how to model the dependencies of sensors. However, existing works basically only consider the micro relationships between sensors, where the sensors are treated equally, and their macroscopic dependencies are neglected. In this paper, we argue to rethink the sensor's dependency modeling from two hierarchies: regional and global perspectives. Particularly, we merge original sensors with high intra-region correlation as a region node to preserve the inter-region dependency. Then, we generate representative and common spatio-temporal patterns as global nodes to reflect a global dependency between sensors and provide auxiliary information for spatio-temporal dependency learning. In pursuit of the generality and reality of node representations, we incorporate a Meta GCN to calibrate the regional and global nodes in the physical data space. Furthermore, we devise the cross-hierarchy graph convolution to propagate information from different hierarchies. In a nutshell, we propose a Hierarchical Information Enhanced Spatio-Temporal prediction method, HIEST, to create and utilize the regional dependency and common spatiotemporal patterns. Extensive experiments have verified the leading performance of our HIEST against state-of-the-art baselines. We publicize the code to ease reproducibility 1 .

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Rethinking Sensors Modeling: Hierarchical Information Enhanced Traffic Forecasting

Ma,

Zhang,

Zhao

et al. 2023

Proceedings of the 32nd ACM International Conference on Information and Knowledge Management

View full text Add to dashboard Cite

show abstract

“…Link prediction (Kumar et al 2020;Daud et al 2020) is a fundamental task in network analysis that aims to predict missing or potential links in a network. It plays a crucial role in various fields, including (i) social network analysis (Kossinets and Watts 2006): suggesting new friendships; (ii) recommender systems (Vahidi Farashah et al 2021): recommending relevant items or products to users; (iii) biological networks (Cos ¸kun and Koyutürk 2021): predicting protein-protein interactions; and (iv) pandemic forecasting (Ma et al 2022): predicting the spread of infectious diseases. The objective of link prediction is to infer the likelihood of a link between two nodes in a network based on the observed network structural features and, if available, node attribute features.…”

Section: Introductionmentioning

confidence: 99%

Link Prediction in Multilayer Networks via Cross-Network Embedding

Ren,

Ding,

et al. 2024

AAAI

View full text Add to dashboard Cite

Link prediction is a fundamental task in network analysis, with the objective of predicting missing or potential links. While existing studies have mainly concentrated on single networks, it is worth noting that numerous real-world networks exhibit interconnectedness. For example, individuals often register on various social media platforms to access diverse services, such as chatting, tweeting, blogging, and rating movies. These platforms share a subset of users and are termed multilayer networks. The interlayer links in such networks hold valuable information that provides more comprehensive insights into the network structure. To effectively exploit this complementary information and enhance link prediction in the target network, we propose a novel cross-network embedding method. This method aims to represent different networks in a shared latent space, preserving proximity within single networks as well as consistency across multilayer networks. Specifically, nodes can aggregate messages from aligned nodes in other layers. Extensive experiments conducted on real-world datasets demonstrate the superior performance of our proposed method for link prediction in multilayer networks.

show abstract