Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism

Yijun Feng; Danfeng Zhu; Mu Li; Xiangdong Wu

doi:10.1109/CCIS59572.2023.10262823

Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism

Yijun Feng^*
, Danfeng Zhu
, Mu Li
, Xiangdong Wu

^*Corresponding author for this work

School of Computer Science and Engineering

Beihang University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Recently, urban transportation problems have become increasingly common, leading to the emergence of Intelligent Transportation Systems (ITS). Estimating travel time is an important component of ITS that can help people accurately estimate arrival times for transportation and deliveries. This article presents a deep learning model called Attention-ODE Travel Time Estimation (AODE-TTE) based on self-Attention mechanisms and neural ordinary differential equations to perform travel time estimation tasks. The AODE-TTE model has excellent ability to capture correlations in long sequence data and can predict travel times for a journey trajectory with high accuracy. Compared to traditional deep neural networks used for travel time prediction, AODE-TTE has fewer parameters making it easier to train and use while also avoiding over-fitting issues. Experimental results show that AODE-TTE outperforms classical models in terms of estimation accuracy and robustness when dealing with different-sized datasets while maintaining good predictive performance.

Original language	English
Title of host publication	Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023
Editors	Xuegong Zhang, Mengqi Zhou, Weining Wang, Wenbai Chen, Yaru Zou, Yanna Liu
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	217-224
Number of pages	8
ISBN (Electronic)	9798350304428
DOIs	https://doi.org/10.1109/CCIS59572.2023.10262823
State	Published - 2023
Event	9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023 - Dali, China Duration: 12 Apr 2023 → 13 Apr 2023

Publication series

Name	Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023

Conference

Conference	9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023
Country/Territory	China
City	Dali
Period	12/04/23 → 13/04/23

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 11 Sustainable Cities and Communities

Keywords

Neural Ordinary Differential Equations
Self-Attention
Travel Time Estimation

Access to Document

10.1109/CCIS59572.2023.10262823

Cite this

Feng, Y., Zhu, D., Li, M., & Wu, X. (2023). Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism. In X. Zhang, M. Zhou, W. Wang, W. Chen, Y. Zou, & Y. Liu (Eds.), Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023 (pp. 217-224). (Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CCIS59572.2023.10262823

Feng, Yijun ; Zhu, Danfeng ; Li, Mu et al. / Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism. Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023. editor / Xuegong Zhang ; Mengqi Zhou ; Weining Wang ; Wenbai Chen ; Yaru Zou ; Yanna Liu. Institute of Electrical and Electronics Engineers Inc., 2023. pp. 217-224 (Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023).

@inproceedings{48b32694832340feadefae658dcca92d,

title = "Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism",

abstract = "Recently, urban transportation problems have become increasingly common, leading to the emergence of Intelligent Transportation Systems (ITS). Estimating travel time is an important component of ITS that can help people accurately estimate arrival times for transportation and deliveries. This article presents a deep learning model called Attention-ODE Travel Time Estimation (AODE-TTE) based on self-Attention mechanisms and neural ordinary differential equations to perform travel time estimation tasks. The AODE-TTE model has excellent ability to capture correlations in long sequence data and can predict travel times for a journey trajectory with high accuracy. Compared to traditional deep neural networks used for travel time prediction, AODE-TTE has fewer parameters making it easier to train and use while also avoiding over-fitting issues. Experimental results show that AODE-TTE outperforms classical models in terms of estimation accuracy and robustness when dealing with different-sized datasets while maintaining good predictive performance.",

keywords = "Neural Ordinary Differential Equations, Self-Attention, Travel Time Estimation",

author = "Yijun Feng and Danfeng Zhu and Mu Li and Xiangdong Wu",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023 ; Conference date: 12-04-2023 Through 13-04-2023",

year = "2023",

doi = "10.1109/CCIS59572.2023.10262823",

language = "英语",

series = "Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "217--224",

editor = "Xuegong Zhang and Mengqi Zhou and Weining Wang and Wenbai Chen and Yaru Zou and Yanna Liu",

booktitle = "Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023",

address = "美国",

}

Feng, Y, Zhu, D, Li, M & Wu, X 2023, Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism. in X Zhang, M Zhou, W Wang, W Chen, Y Zou & Y Liu (eds), Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023. Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023, Institute of Electrical and Electronics Engineers Inc., pp. 217-224, 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023, Dali, China, 12/04/23. https://doi.org/10.1109/CCIS59572.2023.10262823

Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism. / Feng, Yijun; Zhu, Danfeng; Li, Mu et al.
Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023. ed. / Xuegong Zhang; Mengqi Zhou; Weining Wang; Wenbai Chen; Yaru Zou; Yanna Liu. Institute of Electrical and Electronics Engineers Inc., 2023. p. 217-224 (Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism

AU - Feng, Yijun

AU - Zhu, Danfeng

AU - Li, Mu

AU - Wu, Xiangdong

PY - 2023

Y1 - 2023

N2 - Recently, urban transportation problems have become increasingly common, leading to the emergence of Intelligent Transportation Systems (ITS). Estimating travel time is an important component of ITS that can help people accurately estimate arrival times for transportation and deliveries. This article presents a deep learning model called Attention-ODE Travel Time Estimation (AODE-TTE) based on self-Attention mechanisms and neural ordinary differential equations to perform travel time estimation tasks. The AODE-TTE model has excellent ability to capture correlations in long sequence data and can predict travel times for a journey trajectory with high accuracy. Compared to traditional deep neural networks used for travel time prediction, AODE-TTE has fewer parameters making it easier to train and use while also avoiding over-fitting issues. Experimental results show that AODE-TTE outperforms classical models in terms of estimation accuracy and robustness when dealing with different-sized datasets while maintaining good predictive performance.

AB - Recently, urban transportation problems have become increasingly common, leading to the emergence of Intelligent Transportation Systems (ITS). Estimating travel time is an important component of ITS that can help people accurately estimate arrival times for transportation and deliveries. This article presents a deep learning model called Attention-ODE Travel Time Estimation (AODE-TTE) based on self-Attention mechanisms and neural ordinary differential equations to perform travel time estimation tasks. The AODE-TTE model has excellent ability to capture correlations in long sequence data and can predict travel times for a journey trajectory with high accuracy. Compared to traditional deep neural networks used for travel time prediction, AODE-TTE has fewer parameters making it easier to train and use while also avoiding over-fitting issues. Experimental results show that AODE-TTE outperforms classical models in terms of estimation accuracy and robustness when dealing with different-sized datasets while maintaining good predictive performance.

KW - Neural Ordinary Differential Equations

KW - Self-Attention

KW - Travel Time Estimation

UR - https://www.scopus.com/pages/publications/85174976435

U2 - 10.1109/CCIS59572.2023.10262823

DO - 10.1109/CCIS59572.2023.10262823

M3 - 会议稿件

AN - SCOPUS:85174976435

T3 - Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023

SP - 217

EP - 224

BT - Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023

A2 - Zhang, Xuegong

A2 - Zhou, Mengqi

A2 - Wang, Weining

A2 - Chen, Wenbai

A2 - Zou, Yaru

A2 - Liu, Yanna

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023

Y2 - 12 April 2023 through 13 April 2023

ER -

Feng Y, Zhu D, Li M, Wu X. Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism. In Zhang X, Zhou M, Wang W, Chen W, Zou Y, Liu Y, editors, Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023. Institute of Electrical and Electronics Engineers Inc. 2023. p. 217-224. (Proceeding of 2023 9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023). doi: 10.1109/CCIS59572.2023.10262823

Travel Time Estimation Neural Differential Equation Model based on Self-Attention Mechanism

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this