AirFlow: Context Preserving and Multi-Rate State Modeling for Air Quality Forecasting
2026-08-10 • Artificial Intelligence
Artificial IntelligenceComputational Engineering, Finance, and ScienceMachine Learning
AI summaryⓘ
The authors developed AirFlow, a new system to better predict air pollution levels, which is important for health and city planning. Unlike previous methods that treat all pollutants the same way, their approach uses a special normalization method tailored to each pollutant's unique behavior. They also created a dual-stream model to capture patterns at multiple time scales and improve how information is shared between pollutants. Tests on real data from several cities showed AirFlow predicts pollution more accurately than other leading methods while using less computing power.
air quality forecastingnormalizationautocorrelationdistribution driftstate space modelcross-attentionmultivariate time seriespollutant channelsroot mean square errorcomputational efficiency
Authors
Fan Yang, Nan Chen, Yijie Dong, Yuchen Zhang, Wei Zhang
Abstract
Accurate air quality forecasting is essential for public health and urban environmental management, but remains challenging because pollutant channels differ in periodicity and distribution drift, while their concentration trajectories contain both multi-scale dependencies and rapid changes. Recent methods have improved spatial dependency learning and meteorological covariate modeling. However, pollutant channels are still passed through the same normalization rule and temporal backbone, using a shared latent representation for channel-specific distributions and changes at different rates. To address this limitation, we propose AirFlow, a pollutant-aware dual-stream framework that operates on station multivariate observations without additional graph propagation or predefined signal decomposition. Specifically, AirFlow designs two novel blocks: (1) a statistic-guided normalization routing mechanism that selects a normalization path for each pollutant according to its 24-hour autocorrelation and distribution drift; and (2) a hierarchical dual-stream state model that combines multi-scale state space propagation with learnable response coefficients, where gated bidirectional cross-attention exchanges information and adaptively fuses the resulting representations. Experiments on real-world data from multiple cities show that AirFlow achieves the best performance in 34 of 36 metrics comparisons, with reductions of up to 11.11% root mean square error over the state-of-the-art baseline. AirFlow also requires only 0.0483M parameters and 0.0215G FLOPs, achieving high forecasting accuracy with low computational overhead.