The Quo Vadis submission at Traffic4cast 2019
This work addresses traffic forecasting for urban planning, but it is incremental as it builds on existing methods with minor enhancements.
The paper tackled traffic flow prediction in the Traffic4cast competition by using a temporal regression model with spatio-temporal biases, achieving a mean squared error of 9.47×10^-3 and placing eighth in the competition.
We describe the submission of the Quo Vadis team to the Traffic4cast competition, which was organized as part of the NeurIPS 2019 series of challenges. Our system consists of a temporal regression module, implemented as $1\times1$ 2d convolutions, augmented with spatio-temporal biases. We have found that using biases is a straightforward and efficient way to include seasonal patterns and to improve the performance of the temporal regression model. Our implementation obtains a mean squared error of $9.47\times 10^{-3}$ on the test data, placing us on the eight place team-wise. We also present our attempts at incorporating spatial correlations into the model; however, contrary to our expectations, adding this type of auxiliary information did not benefit the main system. Our code is available at https://github.com/danoneata/traffic4cast.