Spatial-Temporal Graph Learning with Adversarial Contrastive Adaptation

TL;DR

GraphST is a novel adversarial contrastive learning model for spatial-temporal graph data that enhances robustness and representation quality in urban sensing tasks by effectively handling data noise and incompleteness.

Contribution

The paper introduces GraphST, a self-supervised adversarial contrastive learning framework that improves spatial-temporal graph representations by adaptively identifying hard samples and modeling inter-view dependencies.

Findings

Outperforms existing models on real-world spatial-temporal prediction tasks.

Enhances robustness against noisy and incomplete data.

Effectively captures inter-view dependencies in region representations.

Abstract

Spatial-temporal graph learning has emerged as a promising solution for modeling structured spatial-temporal data and learning region representations for various urban sensing tasks such as crime forecasting and traffic flow prediction. However, most existing models are vulnerable to the quality of the generated region graph due to the inaccurate graph-structured information aggregation schema. The ubiquitous spatial-temporal data noise and incompleteness in real-life scenarios pose challenges in generating high-quality region representations. To address this challenge, we propose a new spatial-temporal graph learning model (GraphST) for enabling effective self-supervised learning. Our proposed model is an adversarial contrastive learning paradigm that automates the distillation of crucial multi-view self-supervised information for robust spatial-temporal graph augmentation. We empower GraphST to adaptively identify hard samples for better self-supervision, enhancing the representation discrimination ability and robustness. In addition, we introduce a cross-view contrastive learning paradigm to model the inter-dependencies across view-specific region representations and preserve underlying relation heterogeneity. We demonstrate the superiority of our proposed GraphST method in various spatial-temporal prediction tasks on real-life datasets. We release our model implementation via the link: \url{https://github.com/HKUDS/GraphST}.