GT-CausIn: a novel causal-based insight for traffic prediction

Ting GAO, Rodrigo Kappes Marques, Lei Yu

Keywords: spatiotemporal forecasting, causal discovery, graph neural networks

Abstract: Traffic forecasting is an important issue of spatiotemporal series prediction. Among different methods, graph neural networks have achieved so far the most promising results, learning relations between graph nodes then becomes a crucial task. However, improvement space is very limited when these relations are learned in a node-to-node manner. The challenge stems from (1) obscure temporal dependencies between different stations, (2) difficulties in defining variables beyond the node level, and (3) no ready-made method to validate the learned relations. To confront these challenges, we define legitimate traffic variables to discover the causal structure of the traffic network. The causal relation is carefully checked with statistic tools and case analysis. We then present a novel model named Graph Spatial-Temporal Network Based on Causal Insight (GT-CausIn), where graph diffusion layers and temporal convolutional network (TCN) layers are integrated with causal knowledge to capture dependencies in spatiotemporal space. Experiments are carried out on two real-world traffic datasets: PEMS-BAY and METR-LA, which show that GT-CausIn significantly outperforms the state-of-the-art models.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning

TL;DR: A model fusing causal knowledge, space dependency and temporal dependency is proposed in this work.