Skip to main content
QUICK REVIEW

[论文解读] Automated Dilated Spatio-Temporal Synchronous Graph Modeling for Traffic Prediction

Guangyin Jin, Fuxian Li|arXiv (Cornell University)|Jul 22, 2022
Traffic Prediction and Management Techniques被引用 8
一句话总结

本文提出 Auto-DSTSGN,一种自动扩张时空同步图网络,通过扩张卷积联合学习动态图结构并捕捉长短期时空依赖。在四个真实世界交通预测数据集上,其性能相比最先进方法提升约10%,同时保持高效性与低GPU占用率。

ABSTRACT

Accurate traffic prediction is a challenging task in intelligent transportation systems because of the complex spatio-temporal dependencies in transportation networks. Many existing works utilize sophisticated temporal modeling approaches to incorporate with graph convolution networks (GCNs) for capturing short-term and long-term spatio-temporal dependencies. However, these separated modules with complicated designs could restrict effectiveness and efficiency of spatio-temporal representation learning. Furthermore, most previous works adopt the fixed graph construction methods to characterize the global spatio-temporal relations, which limits the learning capability of the model for different time periods and even different data scenarios. To overcome these limitations, we propose an automated dilated spatio-temporal synchronous graph network, named Auto-DSTSGN for traffic prediction. Specifically, we design an automated dilated spatio-temporal synchronous graph (Auto-DSTSG) module to capture the short-term and long-term spatio-temporal correlations by stacking deeper layers with dilation factors in an increasing order. Further, we propose a graph structure search approach to automatically construct the spatio-temporal synchronous graph that can adapt to different data scenarios. Extensive experiments on four real-world datasets demonstrate that our model can achieve about 10% improvements compared with the state-of-art methods. Source codes are available at https://github.com/jinguangyin/Auto-DSTSGN.

研究动机与目标

  • 为解决固定、手工设计的时空图在捕捉不同时段动态交通依赖方面的局限性。
  • 克服现有基于GCN的交通预测模型中独立的时间与空间建模模块效率低下且表达能力有限的问题。
  • 通过自动图结构搜索实现图结构与时空表征的端到端联合学习。
  • 在保持计算效率与低内存占用的前提下提升预测准确率,以支持工业部署。

提出的方法

  • 提出一种自动扩张时空同步图(Auto-DSTSG)模块,通过堆叠扩张因子递增的层来捕捉长短期依赖。
  • 引入可微图结构搜索(GSS)机制,自动学习针对不同数据输入与时间周期的最优时空同步图(STSG)。
  • 构建两类图:用于节点邻近关系的空间图(SGs)与用于节点间模式相似性的时间图(TGs)。
  • 使用可学习掩码与可微松弛技术,实现图拓扑在训练过程中基于梯度的优化。
  • 在学习到的STSG上应用图卷积网络(GCNs),以跨节点与时间步聚合消息。
  • 在STSG框架内集成扩张卷积,以扩展感受野而不增加模型深度。
Figure 1 : Example of spatial and temporal graph construction and spatio-temporal dependencies in a network. The blue dotted arrows, green dotted arrows and red dotted arrows respectively represents the cross-time self-connection, cross-time geo-spatial relations and cross-time pattern similarities.
Figure 1 : Example of spatial and temporal graph construction and spatio-temporal dependencies in a network. The blue dotted arrows, green dotted arrows and red dotted arrows respectively represents the cross-time self-connection, cross-time geo-spatial relations and cross-time pattern similarities.

实验结果

研究问题

  • RQ1是否能够通过端到端可训练框架,联合优化图结构与时空表征学习以实现交通预测?
  • RQ2与固定或自适应图方法相比,自动图结构搜索是否能提升预测准确率?
  • RQ3所提出的扩张时空同步图设计是否能有效捕捉短期与长期依赖?
  • RQ4与最先进基线相比,该模型的效率与GPU显存占用如何?
  • RQ5隐藏维度与图卷积跳数的最优配置为何,才能实现鲁棒性能?

主要发现

  • 该模型在四个真实世界交通数据集上相比最先进方法实现约10%的相对准确率提升。
  • 在PEMS04上学习到的STSG包含更多时间图(TGS),因其时间连接的平均度更高;而PEMS08则包含更多空间图(SGs),反映出对数据特性的结构自适应。
  • 搜索阶段的GPU占用率显著低于训练阶段,表明图结构搜索比神经架构搜索更高效。
  • 该模型保持低GPU占用率与低时间消耗,相比STFGNN与STGODE在效率上表现更优,尤其得益于扩张卷积降低了模型复杂度。
  • 在PEMS04上,隐藏维度为48、最大跳数为2时性能最优;在PEMS08上,隐藏维度为48、最大跳数为3时性能最优;更高维度或跳数则出现过拟合。
  • 学习到的STSG可视化结果表明,模型能根据数据特征自适应调整图拓扑,强时空相关区域具有更高连通性。
Figure 2 : Spatio-temporal synchronous graph construction in previous works. (a) is the adjacency matrix in STSGCN and (b) is the adjacency matrix in STFGNN. $SG^{t_{i}}$ and $TG^{t_{i}}$ respectively denote the spatial graph and temporal graph at time step $i$ . $TC^{t_{i}\to t_{j}}$ describes the
Figure 2 : Spatio-temporal synchronous graph construction in previous works. (a) is the adjacency matrix in STSGCN and (b) is the adjacency matrix in STFGNN. $SG^{t_{i}}$ and $TG^{t_{i}}$ respectively denote the spatial graph and temporal graph at time step $i$ . $TC^{t_{i}\to t_{j}}$ describes the

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。