ETA Prediction with Graph Neural Networks in Google Maps

被引:73
|
作者
Derrow-Pinion, Austin [1 ]
She, Jennifer [1 ]
Wong, David [1 ,2 ]
Lange, Oliver [3 ]
Hester, Todd [1 ,4 ]
Perez, Luis [1 ,5 ]
Nunkesser, Marc [3 ]
Lee, Seongjae [3 ]
Guo, Xueying [3 ]
Wiltshire, Brett [1 ]
Battaglia, Peter W. [1 ]
Gupta, Vishal [1 ]
Li, Ang [1 ]
Xu, Zhongwen [1 ,6 ]
Sanchez-Gonzalez, Alvaro [1 ]
Li, Yujia [1 ]
Velickovic, Petar [1 ]
机构
[1] DeepMind, London, England
[2] Waymo, Mountain View, CA USA
[3] Google, New York, NY USA
[4] Amazon, Santa Fe, NM USA
[5] Facebook AI, Milwaukee, WI USA
[6] Sea AI Lab, Singapore, Singapore
关键词
Graph neural networks; MetaGradients; Google Maps;
D O I
10.1145/3459637.3481916
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Travel-time prediction constitutes a task of high importance in transportation networks, with web mapping services like Google Maps regularly serving vast quantities of travel time queries from users and enterprises alike. Further, such a task requires accounting for complex spatiotemporal interactions (modelling both the topological properties of the road network and anticipating events-such as rush hours-that may occur in the future). Hence, it is an ideal target for graph representation learning at scale. Here we present a graph neural network estimator for estimated time of arrival (ETA) which we have deployed in production at Google Maps. While our main architecture consists of standard GNN building blocks, we further detail the usage of training schedule methods such as MetaGradients in order to make our model robust and production-ready. We also provide prescriptive studies: ablating on various architectural decisions and training regimes, and qualitative analyses on real-world situations where our model provides a competitive edge. Our GNN proved powerful when deployed, significantly reducing negative ETA outcomes in several regions compared to the previous production baseline (40+% in cities like Sydney).
引用
收藏
页码:3767 / 3776
页数:10
相关论文
共 50 条
  • [1] Prediction of steel nanohardness by using graph neural networks on surface polycrystallinity maps
    Karimi, Kamran
    Salmenjoki, Henri
    Mulewska, Katarzyna
    Kurpaska, Lukasz
    Kosinska, Anna
    Alava, Mikko J.
    Papanikolaou, Stefanos
    [J]. SCRIPTA MATERIALIA, 2023, 234
  • [2] On the Prediction Instability of Graph Neural Networks
    Klabunde, Max
    Lemmerich, Florian
    [J]. MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES, ECML PKDD 2022, PT III, 2023, 13715 : 187 - 202
  • [3] Invertible Neural Networks for Graph Prediction
    Xu, Chen
    Cheng, Xiuyuan
    Xie, Yao
    [J]. IEEE Journal on Selected Areas in Information Theory, 2022, 3 (03): : 454 - 467
  • [4] Motif Prediction with Graph Neural Networks
    Besta, Maciej
    Grob, Raphael
    Miglioli, Cesare
    Bernold, Nicola
    Kwasniewski, Grzegorz
    Gjini, Gabriel
    Kanakagiri, Raghavendra
    Ashkboos, Saleh
    Gianinazzi, Lukas
    Dryden, Nikoli
    Hoefler, Torsten
    [J]. PROCEEDINGS OF THE 28TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, KDD 2022, 2022, : 35 - 45
  • [5] Algorithm for Graph Building Based on Google Maps and Google Earth
    Toman, J.
    Olszewska, J. I.
    [J]. 2014 IEEE 15TH INTERNATIONAL SYMPOSIUM ON COMPUTATIONAL INTELLIGENCE AND INFORMATICS (CINTI), 2014, : 55 - 60
  • [6] Link Prediction Based on Graph Neural Networks
    Zhang, Muhan
    Chen, Yixin
    [J]. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018), 2018, 31
  • [7] Nucleophilicity Prediction Using Graph Neural Networks
    Nie, Wan
    Liu, Deguang
    Li, Shuaicheng
    Yu, Haizhu
    Fu, Yao
    [J]. JOURNAL OF CHEMICAL INFORMATION AND MODELING, 2022, 62 (18) : 4319 - 4328
  • [8] Footfall Prediction Using Graph Neural Networks
    Boz, Hasan Alp
    Bahrami, Mohsen
    Balcisoy, Selim
    Pentland, Alex
    [J]. 2023 31ST SIGNAL PROCESSING AND COMMUNICATIONS APPLICATIONS CONFERENCE, SIU, 2023,
  • [9] Equivariant Graph Neural Networks for Toxicity Prediction
    Cremer, Julian
    Medrano Sandonas, Leonardo
    Tkatchenko, Alexandre
    Clevert, Djork-Arne
    De Fabritiis, Gianni
    [J]. CHEMICAL RESEARCH IN TOXICOLOGY, 2023, 36 (10) : 1561 - 1573
  • [10] Disease Prediction via Graph Neural Networks
    Sun, Zhenchao
    Yin, Hongzhi
    Chen, Hongxu
    Chen, Tong
    Cui, Lizhen
    Yang, Fan
    [J]. IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, 2021, 25 (03) : 818 - 826