DMGAN: Dynamic Multi-Hop Graph Attention Network for Traffic Forecasting

Li, Rui; Zhang, Fan; Li, Tong; Zhang, Ning; Zhang, Tingting

doi:10.1109/tkde.2022.3221316

“…We conduct all experiments on a public online platform Kaggle 1 , which provides a virtual NVDIA TESLA P100 GPU card with max 15.9 GB GPU memory to per user. Our model is implemented with the PyTorch framework and the source code and data are publicly available 2 . 2) Configurations: UrbanGCL is trained by using the Adam optimizer with an initial learning rate of 0.001.…”

Section: B Experimental Settings 1) Comuting Infrastructurementioning

confidence: 99%

“…W ITH the rapid pace of urbanization over the past decade, Intelligent Transportation Systems(ITS) have become increasingly indispensable for the evolution of smart cities. Meanwhile, the collection and accumulation of largescale traffic data from road sensors, Taxi GPS trajectories and bicycle rental records, have played a crucial role in traffic condition research and urban traffic flow forecasting [1], [2]. As an important task in ITS, accurate urban traffic flow forecasting not only facilitates the decision-making process for traffic management and strengthens public safety, but also contributes to the improvement of the travel experience for citizens [3], [4].…”

Section: Introductionmentioning

confidence: 99%

Spatial-Temporal Graph Contrastive Learning for Urban Traffic Flow Forecasting

Pan,

Ren,

Li

2023

Preprint

0

View full text Add to dashboard Cite

<p>Graph Neural Networks(GNNs) integrating contrastive learning have attracted growing attentions in urban traffic flow forecasting. However, most existing graph contrastive learning works donot perform well in capturing local-global spatial dependencies and designing contrastive learning scheme in both spatial and temporal dimensions. We argue that these works cannot well extract the spatial-temporal features and are easily affected by data noise. In light of these challenges, this paper proposes an innovative Urban Spatial-Temporal Graph Contrastive Learning framework(UrbanGCL) to improve the accuracy of urban traffic flow forecasting. Specifically, UrbanGCL proposes feature-level and topology-level data augmentation to solve data noise and incompleteness, learn both local and global topology features. Afterwards, the augmented traffic feature matrices and adjacency matrices are fed into a simple yet effective dual-branch network with shared parameters to capture spatialtemporal correlations within traffic sequences. Moreover, we propose the spatial and temporal contrastive learning auxiliary tasks, so as to alleviate the sparsity of supervision signal, and extract the most critical spatial-temporal information. Extensive experimental results on four real-world urban datasets demonstrate that UrbanGCL significantly outperforms other state-ofthe-art methods. The source code of our model is available at https://github.com/panlinnn/UrbanGCL.</p>

show abstract

“…For example, Zeng et al used interactive attention combined with convolution to analyze the temporal tightness and periodicity of crowd flow by using feature fusion to capture different levels of complex correlations (23). Li et al proposed a dynamic graph generated by adapting the spatial attention mechanism for traffic flow prediction by using a predefined graph as a mask (24). Cai et al proposed an attention-based traffic flow prediction method that directly models the temporal relationship between each time step by using transformer's self-attention mechanism (25).…”

Section: Related Workmentioning

confidence: 99%

Research on Traffic Flow Forecasting Based on Dynamic Spatial-Temporal Transformer

Zhang,

Wang,

Zhang

et al. 2023

Transportation Research Record: Journal of the Transportation R

1

0

View full text Add to dashboard Cite

Accurate traffic flow forecasting is crucial for urban traffic control and route planning. Aiming at the difficulty in capturing dynamic spatio-temporal complexity of traffic flow, a dynamic spatio-temporal transformer (DST-Trans) model capable of modeling dynamic correlation of traffic flow is proposed, which consists of gated temporal convolutional network (GTCN), graph convolutional network (GCN), and spatio-temporal transformer (ST-TF). GTCN and GCN are utilized to capture the temporal and spatial characteristics of traffic flow, respectively. ST-TF includes a temporal transformer using temporal gated convolution and temporal multi-head self-attention to capture short-long term temporal features, and spatial transformer using spatial gated graph convolution and spatial multi-head self-attention to capture local-global dynamic spatial features. In addition, to take full advantage of the dynamic and static associations of road networks, multi-graph models of road relationship graph, similarity graph, and adaptive dynamic graph with SGGC are constructed. Experimental results show that the DST-Trans model in this paper shows good prediction performance in short-term (15 min), medium-term (30 min), and long-term (60 min) prediction, outperforming existing state-of-the-art models by up to approximately 7%.

show abstract

“…W ITH the rapid pace of urbanization over the past decade, Intelligent Transportation Systems(ITS) have become increasingly indispensable for the evolution of smart cities. Meanwhile, the collection and accumulation of largescale traffic data from road sensors, Taxi GPS trajectories and bicycle rental records, have played a crucial role in traffic condition research and urban traffic flow forecasting [1], [2]. As an important task in ITS, accurate urban traffic flow forecasting not only facilitates the decision-making process for traffic management and strengthens public safety, but also contributes to the improvement of the travel experience for citizens [3], [4].…”

Section: Introductionmentioning

confidence: 99%

Spatial-Temporal Graph Contrastive Learning for Urban Traffic Flow Forecasting

Pan,

Ren,

Li

2023

Preprint

0

View full text Add to dashboard Cite

<p>Graph Neural Networks(GNNs) integrating contrastive learning have attracted growing attentions in urban traffic flow forecasting. However, most existing graph contrastive learning works donot perform well in capturing local-global spatial dependencies and designing contrastive learning scheme in both spatial and temporal dimensions. We argue that these works cannot well extract the spatial-temporal features and are easily affected by data noise. In light of these challenges, this paper proposes an innovative Urban Spatial-Temporal Graph Contrastive Learning framework(UrbanGCL) to improve the accuracy of urban traffic flow forecasting. Specifically, UrbanGCL proposes feature-level and topology-level data augmentation to solve data noise and incompleteness, learn both local and global topology features. Afterwards, the augmented traffic feature matrices and adjacency matrices are fed into a simple yet effective dual-branch network with shared parameters to capture spatialtemporal correlations within traffic sequences. Moreover, we propose the spatial and temporal contrastive learning auxiliary tasks, so as to alleviate the sparsity of supervision signal, and extract the most critical spatial-temporal information. Extensive experimental results on four real-world urban datasets demonstrate that UrbanGCL significantly outperforms other state-ofthe-art methods. The source code of our model is available at https://github.com/panlinnn/UrbanGCL.</p>

show abstract

DMGAN: Dynamic Multi-Hop Graph Attention Network for Traffic Forecasting

Cited by 8 publications

References 26 publications

Spatial-Temporal Graph Contrastive Learning for Urban Traffic Flow Forecasting

Spatial-Temporal Graph Contrastive Learning for Urban Traffic Flow Forecasting

Research on Traffic Flow Forecasting Based on Dynamic Spatial-Temporal Transformer

Spatial-Temporal Graph Contrastive Learning for Urban Traffic Flow Forecasting

Contact Info

Product

Resources

About