Modeling Spatial Nonstationarity via Deformable Convolutions for Deep Traffic Flow Prediction

  title={Modeling Spatial Nonstationarity via Deformable Convolutions for Deep Traffic Flow Prediction},
  author={Wei Zeng and Chengqiao Lin and Kang Liu and Juncong Lin and Anthony K. H. Tung},
Deep neural networks are being increasingly used for short-term traffic flow prediction. Existing convolution-based approaches typically partition an underlying territory into grid-like spatial units, and employ standard convolutions to learn spatial dependence among the units. However, standard convolutions with fixed geometric structures cannot fully model the nonstationary characteristics of local traffic flows. To overcome the deficiency, we introduce deformable convolution that augments… 

Figures and Tables from this paper


Attention Based Spatial-Temporal Graph Convolutional Networks for Traffic Flow Forecasting
Experiments on two real-world datasets from the Caltrans Performance Measurement System demonstrate that the proposed ASTGCN model outperforms the state-of-the-art baselines.
Deep Spatial–Temporal 3D Convolutional Neural Networks for Traffic Data Forecasting
A novel end-to-end deep learning model, called ST-3DNet, is proposed, which introduces 3D convolutions to automatically capture the correlations of traffic data in both spatial and temporal dimensions and a novel recalibration block is proposed to explicitly quantify the difference of the contributions of the correlations in space.
Deep Spatio-Temporal Residual Networks for Citywide Crowd Flows Prediction
A deep-learning-based approach to collectively forecast the inflow and outflow of crowds in each and every region of a city, using the residual neural network framework to model the temporal closeness, period, and trend properties of crowd traffic.
DeepSTN+: Context-Aware Spatial-Temporal Neural Network for Crowd Flow Prediction in Metropolis
DeepSTN+ is proposed, a deep learning-based convolutional model to predict crowd flows in the metropolis that reduces the error of the crowd flow prediction by approximately 8%∼13% compared with the state-of-the-art baselines.
Deep Irregular Convolutional Residual LSTM for Urban Traffic Passenger Flows Prediction
  • Bowen Du, Hao Peng, +5 authors J. Li
  • Computer Science
    IEEE Transactions on Intelligent Transportation Systems
  • 2020
A deep irregular convolutional residual LSTM network model called DST-ICRL is proposed which significantly outperforms both traditional and deep learning based urban traffic passenger flows prediction methods.
T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction
A novel neural network-based traffic forecasting method, the temporal graph convolutional network (T-GCN) model, which is combined with the graph convolved network (GCN), and the gated recurrent unit (GRU) to capture the spatial and temporal dependences simultaneously.
Flow Prediction in Spatio-Temporal Networks Based on Multitask Deep Learning
A multitask deep-learning framework that simultaneously predicts the node flow and edge flow throughout a spatio-temporal network based on fully convolutional networks is proposed.
Deep Multi-View Spatial-Temporal Network for Taxi Demand Prediction
A Deep Multi-View Spatial-Temporal Network (DMVST-Net) framework to model both spatial and temporal relations is proposed, which demonstrates effectiveness of the approach over state-of-the-art methods.
Traffic Flow Prediction With Big Data: A Deep Learning Approach
A novel deep-learning-based traffic flow prediction method is proposed, which considers the spatial and temporal correlations inherently and is applied for the first time that a deep architecture model is applied using autoencoders as building blocks to represent traffic flow features for prediction.
DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
This work addresses the task of semantic image segmentation with Deep Learning and proposes atrous spatial pyramid pooling (ASPP), which is proposed to robustly segment objects at multiple scales, and improves the localization of object boundaries by combining methods from DCNNs and probabilistic graphical models.