Research

Understanding Urban Economic Status through GNN-based Urban Representation Learning Using Mobility Data 

With growing urban population and urban concentration, various data-driven efforts are being made to achieve sustainable growth to promote equity, inclusion, and well-being. Among abundant urban data, mobility data is a source with rich semantic about urban environments in which social and economic activities are dissolved. In this paper, we employ graph attention network (GAT) to obtain urban representation learning embedding based on taxi trips and subway ridership data in Seoul, South Korea. Our GAT-based region embedding model outperformed all baseline models in predicting the number of employees and housing prices. For the number of employees prediction, our model achieved R-squared value of 0.649 using mobility data only. We also found that increasing the embedding dimensions to stack the elderly and disabled subway user types can further improve the model’s capability in the number of employees and housing prices predictions. Our study results suggest that a transportation network is a key contributor to shaping the economic landscapes of urban regions. Such findings also indicate that understanding people’s activities and movements is essential in achieving sustainable urban growth and promoting equity and inclusion for all. Our research contributes to the growing body of research on the urban region representation learning in understanding the economic impact of transportation systems on urban regions, especially for vulnerable populations. 

Reference:

• Y. Shin, G. Seong, N. Kim, S. Kim, Y. Yoon, "Understanding Urban Economics Status through GNN-based Urban Representation Learning Using Mobility Data," 1st ACM SIGSPATIAL International Workshop on Advances in Urban-AI (UrbanAI 2023), 2023. [Link]

Deep Learning Framework for Vessel Trajectory Prediction Using Auxiliary Tasks and Convolutional Networks

The accurate prediction of vessel trajectories plays a pivotal role in various maritime applications, including route planning, collision avoidance, and maritime traffic management. With the exponential growth in vessel traffic and the increasing complexity of maritime operations, there is a pressing need for reliable and efficient methods to forecast vessel movements. Traditional statistical and machine-learning approaches have limitations in capturing the complex spatial-temporal patterns of vessel movements. Deep learning techniques have emerged as a promising solution due to their ability to handle large-scale datasets and capture nonlinear relationships. This paper proposes a novel deep learning-based vessel trajectory prediction framework for AIS data using Auxiliary tasks and Convolutional encoders (AIS-ACNet). The model leverages Automatic Identification System (AIS) data, including geographical positions, and vessel dynamics such as Speed Over Ground (SOG), and Course Over Ground (COG), for trajectory prediction. The AIS-ACNet employs parallel convolutional encoder networks with feature fusion layers. The model is trained with a learning objective that includes auxiliary tasks such as SOG and COG predictions. This framework enhances the model's representative power of vessel trajectory data leading to a better understanding of vessel dynamics and higher trajectory prediction performance. The proposed framework is evaluated on real-world data from the Port of Houston, Texas, USA, and compared to existing models through extensive experiments and ablation studies. The results demonstrate the effectiveness and superiority of AIS-ACNet in accurately predicting vessel trajectories.

Reference:

• Y. Shin, N. Kim, H. Lee, S.Y. In, M. Hansen, and Y. Yoon*, "Deep Learning Framework for Vessel Trajectory Prediction Using Auxiliary Tasks and Convolutional Networks" (Revision submitted - EAAI).

A Novel Graph Convolutional Networks to Progressively Adapt to Online Data

​​​​In recent years, graph convolutional networks that possess the ability to adapt to the input data have shown promising results in several studies. In most cases, however, the adaptation has been made during the training phases of the models, which inevitably make the models vulnerable to unexpected traffic conditions such as road closure and traffic accident during the testing phases. In this study, we propose a novel traffic forecasting model, Progressive Graph Convolutional Network (PGCN) to make the model adapt to online traffic data. PGCN constructs a set of graphs by calculating learnable similarity measures among the node signals. The architecture of the model is based on Graph WaveNet. When applied to four real-world datasets, PGCN consistently achieves state-of-the-art performance. This result shows that PGCN has the ability to generalize in different study areas by progressively adapting to online data.

Reference:

• Y. Shin, Y.Yoon*, "PGCN: Progressive Graph Convolutional Networks for Spatial-Temporal Traffic Forecasting", ArXiv preprint, arXiv:2202.08982, 2022 (Revision submitted -IEEE T-ITS). [Link]

Performance Evaluation of Basic Elements of Deep Neural Network Models for Traffic Forecasting

Since 2014, various studies have proposed deep learning-based models to solve traffic forecasting problems. While earlier approaches have shown a more wide range of implementations, the basic elements of the recent models can be categorized into a few groups - RNN, convolution, and self-attention for temporal feature extraction, and Graph Convolutional Networks (GCN) and Graph Attention Networks (GAT) for spatial feature extraction. However, there has been little effort to assess the characteristics of each element and make an in-depth evaluation. In this study, we thoroughly evaluated the performance and characteristics of basic elements of traffic forecasting models through extensive and multi-faceted experiments on four real-world datasets. The result reveals that there is no single element that is superior to the others in all aspects. Interesting outcomes include that the convolution-based models show more accurate overall performance than the attention-based models, while the attention-based models show more robustness against abnormal conditions.  

References 

• Y. Shin, Y. Yoon*, "Performance Evaluation of Building Blocks of Spatial-Temporal Deep Learning Models for Traffic Forecasting", (Accepted at IEEE Access). [Link]

• Y. Shin, Y. Yoon*, "A Comparative Study on Basic Elements of Deep Learning Models for Spatial-Temporal Traffic Forecasting", AAAI-22 workshop: AI for Transportation, 2022. [Link]

GNN-based Traffic Forecasting Using Various Structural Characteristics 

Traffic forecasting is a research area in transportation engineering that has garnered broad research interest as a key technical enabler of intelligent transportation systems. The recent surge of Graph Convolutional Networks has enabled traffic forecasting models to incorporate the complex relationship among nodes of transportation networks, and further improved the performance of deep learning-based models. However, many studies overlook the structural characteristics (i.e. speed limit, distance, and flow direction) of transportation networks which can bring important information about the networks. In this research, we propose Multi-Weight Traffic Graph Convolutional (MW-TGC) Networks to incorporate the aforementioned features in a single model and to reflect more spatial correlation in the traffic forecasting problem. Experiments on two real-world datasets show that the proposed MW-TGC Network outperforms the state-of-the-art models. 

References 

• Y. Shin, Y. Yoon*, "Incorporating Dynamicity of Transportation Network With Multi-Weight Traffic Graph Convolutional Network for Traffic Forecasting", IEEE Transactions on Intelligent Transportation Systems, 23(3), 2082-2092, 2022. [Link]