Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting

Authors: Yaguang Li, Rose Yu, Cyrus Shahabi, Yan Liu

ICLR 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate the framework on two real-world large scale road network traffic datasets and observe consistent improvement of 12% 15% over state-of-the-art baselines. We conducted extensive experiments on two large-scale real-world datasets, and the proposed approach obtains significant improvement over state-of-the-art baseline methods.
Researcher Affiliation Academia University of Southern California, California Institute of Technology {yaguang, shahabi, yanliu.cs}@usc.edu, rose@caltech.edu
Pseudocode No The paper describes mathematical equations for its models (e.g., DCGRU equations), and provides a system architecture diagram (Figure 2), but does not include any explicit pseudocode blocks or algorithms.
Open Source Code Yes 1The source code is available at https://github.com/liyaguang/DCRNN.
Open Datasets Yes 70% of data is used for training, 20% are used for testing while the remaining 10% for validation.
Dataset Splits Yes 70% of data is used for training, 20% are used for testing while the remaining 10% for validation.
Hardware Specification No The paper mentions using TensorFlow and an Adam optimizer, but it does not specify any hardware details such as CPU, GPU models, or memory specifications used for the experiments.
Software Dependencies No All neural network based approaches are implemented using Tensorflow (Abadi et al., 2016), and trained using the Adam optimizer with learning rate annealing. ARIMAkal: ... model is implemented using the statsmodel python package. VAR: ... model is implemented using the statsmodel python package. (Specific version numbers for TensorFlow or statsmodel are not provided.)
Experiment Setup Yes The best hyperparameters are chosen using the Tree-structured Parzen Estimator (TPE) (Bergstra et al., 2011) on the validation dataset. Detailed parameter settings for DCRNN as well as baselines are available in Appendix E. (Appendix E provides specific details for FNN, FC-LSTM, and DCRNN including hidden layer units, learning rates, epochs, dropout, weight decay, batch size, loss function, K value, and scheduled sampling parameters).