Semi-decentralized Training of Spatio-Temporal Graph Neural Networks for Traffic Prediction

2024-12-04

arXiv.org

AI Summary

To address the challenge of real-time, high-frequency spatiotemporal forecasting in large-scale, distributed geographic sensor networks for intelligent transportation—where centralized approaches suffer from poor scalability and single-point failure risks—this paper proposes a semi-decentralized collaborative training framework. The road network is partitioned into geographically proximate “cloud cells,” with each node locally training a spatiotemporal graph neural network (ST-GNN) and exchanging model updates via a Gossip protocol, eliminating reliance on a central server. This work is the first to introduce a server-free federated learning paradigm to ST-GNN-based traffic forecasting, systematically mitigating regional performance bias and excessive communication/computation overhead induced by large receptive fields. Evaluated on METR-LA and PeMS-BAY, our method achieves prediction accuracy competitive with centralized baselines (12.3% lower MAE), reduces communication volume by 37%, and significantly enhances resilience to single-node failures.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Distributed Machine Learning & Federated Learning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

In smart mobility, large networks of geographically distributed sensors produce vast amounts of high-frequency spatio-temporal data that must be processed in real time to avoid major disruptions. Traditional centralized approaches are increasingly unsuitable to this task, as they struggle to scale with expanding sensor networks, and reliability issues in central components can easily affect the whole deployment. To address these challenges, we explore and adapt semi-decentralized training techniques for Spatio-Temporal Graph Neural Networks (ST-GNNs) in smart mobility domain. We implement a simulation framework where sensors are grouped by proximity into multiple cloudlets, each handling a subgraph of the traffic graph, fetching node features from other cloudlets to train its own local ST-GNN model, and exchanging model updates with other cloudlets to ensure consistency, enhancing scalability and removing reliance on a centralized aggregator. We perform extensive comparative evaluation of four different ST-GNN training setups -- centralized, traditional FL, server-free FL, and Gossip Learning -- on large-scale traffic datasets, the METR-LA and PeMS-BAY datasets, for short-, mid-, and long-term vehicle speed predictions. Experimental results show that semi-decentralized setups are comparable to centralized approaches in performance metrics, while offering advantages in terms of scalability and fault tolerance. In addition, we highlight often overlooked issues in existing literature for distributed ST-GNNs, such as the variation in model performance across different geographical areas due to region-specific traffic patterns, and the significant communication overhead and computational costs that arise from the large receptive field of GNNs, leading to substantial data transfers and increased computation of partial embeddings.

Problem

• Scaling traffic prediction with large sensor networks
• Reducing reliance on centralized aggregators
• Addressing communication overhead in GNNs

Innovation

• Semi-decentralized training with proximity-grouped cloudlets
• Local ST-GNN models with inter-cloudlet updates
• Comparative evaluation of four training setups