Learning to Generate Maps from Trajectories

Authors: Sijie Ruan, Cheng Long, Jie Bao, Chunyang Li, Zisheng Yu, Ruiyuan Li, Yuxuan Liang, Tianfu He, Yu Zheng890-897

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on two real-world trajectory datasets confirm that Deep MG significantly outperforms the state-of-theart methods.
Researcher Affiliation Collaboration 1School of Computer Science and Technology, Xidian University, Xi an, China 2JD Intelligent Cities Research, Beijing, China, 3JD Intelligent Cities Business Unit, JD Digits, Beijing, China 4School of Computer Science and Engineering, Nanyang Technological University, Singapore 5School of Computing, National University of Singapore, Singapore 6School of Computer Science and Technology, Harbin Institute of Technology, Harbin, China
Pseudocode Yes Algorithm 1 Link Generation.
Open Source Code No The paper does not provide any statement or link indicating the availability of its source code.
Open Datasets Yes We use two real-world taxi trajectory datasets, i.e., Taxi BJ and Taxi JN, with different sampling rates for evaluations. Taxi BJ is obtained from T-Drive (Yuan et al. 2010a)
Dataset Splits Yes The number of samples for training, validation, and test are 744, 180, 100 for Taxi BJ, and 1251, 313, 100 for Taxi JN.
Hardware Specification Yes The T2RNet model in geometry translation is implemented by Py Torch, and trained with one NVIDIA Tesla V100 GPU.
Software Dependencies No The paper mentions 'Python' and 'Py Torch' but does not specify version numbers for these or any other key software components, libraries, or solvers.
Experiment Setup Yes During the training phase, we leverage Adam (Kingma and Ba 2014) to perform network training with a learning rate 2e 4 and batch size 8. We also apply a staircase-like schedule by halving the learning rate every 10 epochs. For topology construction, Rlink = 100m, α = 1.4, S = 5 for Taxi BJ and S = 2 for Taxi JN due to different number of trajectories.