Revisiting Spatial-Temporal Similarity: A Deep Learning Framework for Traffic Prediction

Abstract

Traffic prediction has drawn increasing attention in AI research field due to the increasing availability of large-scale traffic data and its importance in the real world. For example, an accurate taxi demand prediction can assist taxi companies in pre-allocating taxis. The key challenge of traffic prediction lies in how to model the complex spatial dependencies and temporal dynamics. Although both factors have been considered in modeling, existing works make strong assumptions about spatial dependence and temporal dynamics, i.e., spatial dependence is stationary in time, and temporal dynamics is strictly periodical. However, in practice the spatial dependence could be dynamic (i.e., changing from time to time), and the temporal dynamics could have some perturbation from one period to another period. In this paper, we make two important observations: (1) the spatial dependencies between locations are dynamic; and (2) the temporal dependency follows daily and weekly pattern but it is not strictly periodic for its dynamic temporal shifting. To address these two issues, we propose a novel Spatial-Temporal Dynamic Network (STDN), in which a flow gating mechanism is introduced to learn the dynamic similarity between locations, and a periodically shifted attention mechanism is designed to handle long-term periodic temporal shifting. To the best of our knowledge, this is the first work that tackle both issues in a unified framework. Our experimental results on real-world traffic datasets verify the effectiveness of the proposed method.

Attention-Based Supply-Demand Prediction for Autonomous Vehicles

“…In this section, we first fix some notations and define the demand-supply prediction problem. We follow previous studies [8] [13] and split the whole city to an X ×Y grid map which consists of X rows and Y columns. define the set of non-overlapping locations L = {l 1 , l 2 , ..., l i , ..., l X×Y }.…”

“…STDN [16]: STDN is a deep learning based model which considers temporal shifting. In this paper, STDN is applied without flow gating.…”

“…To capture both longterm periodic information and temporal shifting, researchers propose to use weighted representation. For example, Yao et al [8] proposed a periodically shifted attention mechanism by taking long-term periodic information and temporal shifting simultaneously. Wang et al [11] used a Softmax Layer to get the weight vector of representation.…”

Learning from Multiple Cities: A Meta-Learning Approach for Spatial-Temporal Prediction

“…holiday) or missing values. These patterns as indicators of region functionality, however, are crucial for spatial-temporal prediction [33,45]. Take the traffic demand prediction as an instance.…”

Cluster-Based LSTM Network for Short-Term Passenger Flow Forecasting in Urban Rail Transit

“…Zhang et al [26] designed an architecture including attention mechanism, GCN and sequence-to-sequence model to conduct multistep speed prediction. After ConvLSTM was firstly introduced [27], CNN and LSTM are often integrated together to perform traffic predictions [28,29]. Recently, generative adversarial network has began to attract researchers' attention and has been applied to traffic time estimation [30].…”