Listening to the City, Attentively: A Spatio-Temporal Attention-Boosted Autoencoder for the Short-Term Flow Prediction Problem

In recent years, studying and predicting mobility patterns in urban environments has become increasingly important as accurate timely information on current future vehicle flows can successfully increase the quality availability of transportation services (e.g., sharing services). However, number incoming outgoing vehicles for different city areas is challenging due to nonlinear spatial temporal dependencies typical patterns. this work, we propose STREED-Net, a novel autoencoder architecture featuring time-distributed convolutions, cascade hierarchical units two distinct attention mechanisms (one one temporal) that effectively captures exploits complex data short-term flow prediction problem. The results thorough experimental analysis using real-life are reported, indicating proposed model improves state-of-the-art task.