Federated Learning for Traffic Flow Prediction with Synthetic Data Augmentation

2024-12-11

arXiv.org

AI Summary

To address privacy constraints, data heterogeneity, and sparsity arising from multi-institutional traffic data silos, this paper proposes FedTPS—a novel federated framework for privacy-preserving synthetic trajectory generation. FedTPS pioneers the federation of diffusion models for spatiotemporal trajectory synthesis and introduces a hybrid spatiotemporal graph neural network that jointly models dynamic road networks and long-range temporal dependencies via integrated temporal and graph attention mechanisms. Built upon the cross-silo federated learning paradigm, FedTPS is rigorously evaluated on large-scale real-world ride-hailing data. Experimental results demonstrate a 12.7% reduction in mean absolute error (MAE) for global traffic flow prediction, significantly outperforming state-of-the-art federated baselines. Key contributions include: (i) the first federated diffusion-based generative mechanism for trajectory synthesis; (ii) a dual-attention collaborative modeling architecture unifying temporal dynamics and topological structure; and (iii) end-to-end privacy preservation with cross-domain data synergy and utility enhancement.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Diffusion Models for Vision

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Deep-learning based traffic prediction models require vast amounts of data to learn embedded spatial and temporal dependencies. The inherent privacy and commercial sensitivity of such data has encouraged a shift towards decentralised data-driven methods, such as Federated Learning (FL). Under a traditional Machine Learning paradigm, traffic flow prediction models can capture spatial and temporal relationships within centralised data. In reality, traffic data is likely distributed across separate data silos owned by multiple stakeholders. In this work, a cross-silo FL setting is motivated to facilitate stakeholder collaboration for optimal traffic flow prediction applications. This work introduces an FL framework, referred to as FedTPS, to generate synthetic data to augment each client's local dataset by training a diffusion-based trajectory generation model through FL. The proposed framework is evaluated on a large-scale real world ride-sharing dataset using various FL methods and Traffic Flow Prediction models, including a novel prediction model we introduce, which leverages Temporal and Graph Attention mechanisms to learn the Spatio-Temporal dependencies embedded within regional traffic flow data. Experimental results show that FedTPS outperforms multiple other FL baselines with respect to global model performance.

Problem

• Addresses data scarcity in traffic flow prediction models
• Enhances privacy and collaboration in decentralized data environments
• Improves global model performance using synthetic data augmentation

Innovation

• Federated Learning for decentralized traffic data analysis
• Synthetic data augmentation using diffusion-based trajectory generation
• Temporal and Graph Attention mechanisms for traffic prediction