Double-Diffusion: Balancing Speed, Accuracy, and Uncertainty in Probabilistic Forecasting for Urban Sensor Networks

2025/06/29 by Arian Prabowo, Dong, Hanlin, Prabowo, Arian +7
Computer Science · Environmental Science · #Advanced Technologies in Various Fields #Air Quality Monitoring and Forecasting #Air Quality and Health Impacts #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2506.23053

openalex publication_date 2025/06/29 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Urban sensor networks need forecasts that are accurate, carry useful uncertainty, and refresh fast enough to act on as new readings arrive. These goals conflict: deterministic models give no distribution, while diffusion forecasters model uncertainty but denoise from pure noise over many steps. We present Double-Diffusion, which integrates a closed-form graph-heat prior into a denoising diffusion model. The prior is a parameter-free low-pass forecast over the sensor graph, and it serves two roles: it is the residual target the model generates, and it conditions the denoiser. The reverse process therefore starts near the prior and denoises a short warm-started chain instead of synthesizing from pure noise; the name denotes this composition, a graph diffusion feeding a denoising diffusion. A compact denoiser, DD-Net, is trained as a Denoising Diffusion Probabilistic Model (DDPM) in the Resfusion warm-start formulation, so generation refines the prior over a short truncated chain rather than synthesizing from pure noise; a graph-spectral read-out of the prior residual sets its switchable spatial filter per domain from training data alone. On four real-world air quality and traffic networks, Double-Diffusion attains the best CRPS of all probabilistic methods on every dataset and stays competitive in point accuracy with the strongest baselines, at a fraction of the sampling cost of from-noise diffusion. The code is available at: https://github.com/teddyicare/Double-Diffusion

Citations

• Elucidated Rolling Diffusion Models for Probabilistic Forecasting of Complex Dynamics
• Air Quality Prediction with Physics-Guided Dual Neural ODEs in Open Systems
• SpecSTG: A Fast Spectral Diffusion Framework for Probabilistic Spatio-Temporal Traffic Forecasting
• Resfusion: Denoising Diffusion Probabilistic Models for Image Restoration Based on Prior Residual Noise
• Towards Unifying Diffusion Models for Probabilistic Spatio-Temporal Graph Learning
• ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting
• PriSTI: A Conditional Diffusion Framework for Spatiotemporal Imputation
• DiffSTG: Probabilistic Spatio-Temporal Graph Forecasting with Denoising Diffusion Models
• AirFormer: Predicting Nationwide Air Quality in China with Transformers
• CSDI: Conditional Score-based Diffusion Models for Probabilistic Time Series Imputation
• Quantifying Uncertainty in Deep Spatiotemporal Forecasting
• Pay Attention to MLPs
• Denoising Diffusion Implicit Models
• Adaptive Universal Generalized PageRank Graph Neural Network
• GMAN: A Graph Multi-Attention Network for Traffic Prediction
• GMAN: A Graph Multi-Attention Network for Traffic Prediction
• Revisiting Graph Neural Networks: All We Have is Low-Pass Filters
• Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Neural Ordinary Differential Equations
• DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networks
• Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning
• The emerging field of signal processing on graphs: Extending high-dimensional data analysis to networks and other irregular domains

Related

• AirPhyNet: Harnessing Physics-Guided Neural Networks for Air Quality Prediction
• STeP-Diff: Spatio-Temporal Physics-Informed Diffusion Models for Mobile Fine-Grained Pollution Forecasting
• A Model-driven and Data-driven Fusion Framework for Accurate Air Quality Prediction
• Difference Learning for Air Quality Forecasting Transport Emulation
• A HEART for the environment: Transformer-Based Spatiotemporal Modeling for Air Quality Prediction