Relational Fusion Networks: Graph Convolutional Networks for Road Networks

Tobias Skovgaard Jepsen; Christian S. Jensen; Thomas Dyhre Nielsen

doi:10.1109/TITS.2020.3011799

Relational Fusion Networks: Graph Convolutional Networks for Road Networks

Tobias Skovgaard Jepsen^*, Christian S. Jensen, Thomas Dyhre Nielsen

^*Corresponding author

Research output: Contribution to journal › Journal article › Research › peer-review

13 Downloads (Pure)

Abstract

The application of machine learning techniques inthe setting of road networks holds the potential to facilitate many important intelligent transportation applications. Graph Convolutional Networks (GCNs) are neural networks that are capable of leveraging the structure of a network. However, many implicit assumptions of GCNs do not apply to road networks. We introduce the Relational Fusion Network (RFN), a novel type of GCN designed specifically for road networks. In particular, we propose methods that outperform state-of-the-art GCN architectures by up to 21–40% on two machine learning tasks in road networks. Furthermore, we show that state-of-the-art GCNs may fail to effectively leverage road network structure and may not generalize well to other road networks

Original language	English
Journal	IEEE Transactions on Intelligent Transportation Systems
Pages (from-to)	1-12
Number of pages	12
ISSN	1524-9050
DOIs	https://doi.org/10.1109/TITS.2020.3011799
Publication status	Published - 14 Aug 2020

Keywords

Transportation
Road Network
Machine Learning
Graph Representation Learning
Graph Convolutional Networks

Access to Document

10.1109/TITS.2020.3011799

rfnSubmitted manuscript, 2.84 MB

https://arxiv.org/abs/2006.09030

Fingerprint Dive into the research topics of 'Relational Fusion Networks: Graph Convolutional Networks for Road Networks'. Together they form a unique fingerprint.

1 Finished

DiCyPS: Center for Data-Intensive Cyber-Physical Systems
Agerholm, N.
01/01/2015 → 31/12/2020
Project: Research

2 Article in proceeding

Scalable Unsupervised Multi-Criteria Trajectory Segmentation and Driving Preference Mining
Barth, F., Funke, S., Skovgaard Jepsen, T. & Proissl, C., 3 Nov 2020, BIGSPATIAL '20: Proceedings of the 9th ACM SIGSPATIAL International Workshop on Analytics for Big Geospatial Data. Chandola, V., Vatsavai, R. R. & Shashidharan, A. (eds.). Association for Computing Machinery, p. 1-10 10 p. 6
Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

File
2 Downloads (Pure)
Graph Convolutional Networks for Road Networks
Skovgaard Jepsen, T., Jensen, C. S. & Nielsen, T. D., 5 Nov 2019, Proceedings of the 27th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems. Banaei-Kashani, F., Trajcevski, G., Guting, R. H., Kulik, L. & Newsam, S. (eds.). Association for Computing Machinery, p. 460-463 4 p.
Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

Open Access
File
2 Citations (Scopus)

152 Downloads (Pure)

Cite this

@article{4cf9ab4a9aa4442d94cb67005afe03ca,

title = "Relational Fusion Networks: Graph Convolutional Networks for Road Networks",

abstract = "The application of machine learning techniques inthe setting of road networks holds the potential to facilitate many important intelligent transportation applications. Graph Convolutional Networks (GCNs) are neural networks that are capable of leveraging the structure of a network. However, many implicit assumptions of GCNs do not apply to road networks. We introduce the Relational Fusion Network (RFN), a novel type of GCN designed specifically for road networks. In particular, we propose methods that outperform state-of-the-art GCN architectures by up to 21–40% on two machine learning tasks in road networks. Furthermore, we show that state-of-the-art GCNs may fail to effectively leverage road network structure and may not generalize well to other road networks",

keywords = "Transportation, Road Network, Machine Learning, Graph Representation Learning, Graph Convolutional Networks",

author = "{Skovgaard Jepsen}, Tobias and Jensen, {Christian S.} and Nielsen, {Thomas Dyhre}",

year = "2020",

month = aug,

day = "14",

doi = "10.1109/TITS.2020.3011799",

language = "English",

pages = "1--12",

journal = "I E E E Transactions on Intelligent Transportation Systems",

issn = "1524-9050",

publisher = "IEEE",

}

TY - JOUR

T1 - Relational Fusion Networks: Graph Convolutional Networks for Road Networks

AU - Skovgaard Jepsen, Tobias

AU - Jensen, Christian S.

AU - Nielsen, Thomas Dyhre

PY - 2020/8/14

Y1 - 2020/8/14

N2 - The application of machine learning techniques inthe setting of road networks holds the potential to facilitate many important intelligent transportation applications. Graph Convolutional Networks (GCNs) are neural networks that are capable of leveraging the structure of a network. However, many implicit assumptions of GCNs do not apply to road networks. We introduce the Relational Fusion Network (RFN), a novel type of GCN designed specifically for road networks. In particular, we propose methods that outperform state-of-the-art GCN architectures by up to 21–40% on two machine learning tasks in road networks. Furthermore, we show that state-of-the-art GCNs may fail to effectively leverage road network structure and may not generalize well to other road networks

AB - The application of machine learning techniques inthe setting of road networks holds the potential to facilitate many important intelligent transportation applications. Graph Convolutional Networks (GCNs) are neural networks that are capable of leveraging the structure of a network. However, many implicit assumptions of GCNs do not apply to road networks. We introduce the Relational Fusion Network (RFN), a novel type of GCN designed specifically for road networks. In particular, we propose methods that outperform state-of-the-art GCN architectures by up to 21–40% on two machine learning tasks in road networks. Furthermore, we show that state-of-the-art GCNs may fail to effectively leverage road network structure and may not generalize well to other road networks

KW - Transportation

KW - Road Network

KW - Machine Learning

KW - Graph Representation Learning

KW - Graph Convolutional Networks

U2 - 10.1109/TITS.2020.3011799

DO - 10.1109/TITS.2020.3011799

M3 - Journal article

SP - 1

EP - 12

JO - I E E E Transactions on Intelligent Transportation Systems

JF - I E E E Transactions on Intelligent Transportation Systems

SN - 1524-9050

ER -

Relational Fusion Networks: Graph Convolutional Networks for Road Networks

Abstract

Keywords

Access to Document

DiCyPS: Center for Data-Intensive Cyber-Physical Systems

Scalable Unsupervised Multi-Criteria Trajectory Segmentation and Driving Preference Mining

Graph Convolutional Networks for Road Networks

Cite this