This paper proposes AirFlow, a pollutant-aware dual-stream framework for accurate air quality forecasting using a statistic-guided normalization routing mechanism and a hierarchical dual-stream state model.
Introduces a pollutant-aware dual-stream framework for air quality forecasting using a statistic-guided normalization routing mechanism and a hierarchical dual-stream state model.
Keywords
Before reading this…
Applications
To understand this paper, make sure you know these concepts first:
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.
Transformer-based Diffusion models for Hydrological Time Series Probabilistic Imputation and Forecas…
This paper investigates the application of transformer-based diffusion models fo…
COGENT: Continuous Graph Emulators with Neural Ordinary Differential Equations for Long-Term Physica…
COGENT is a continuous graph emulator with Neural Ordinary Differential Equation…
A Comparative Study of Deep Learning Architectures for Multi-Horizon Behavioural Forecasting for Mob…
This paper compares and evaluates six deep learning architectures for forecastin…
Co-Evolving Acting and Forecasting for Language Agents
The paper proposes a new method, ame, to improve the performance of language ag…
CLUIE: Clustering-Aware Recurrent Propagation with Local Structural Compensation for Underwater Imag…
A new framework, CRWKV, is proposed for underwater image enhancement using a Clu…
WiFi based Multi user Activity Recognition via User Conditioned Spatial Attention
The paper proposes a method for WiFi-based human activity recognition in multi-u…
Cross-Market Meta-Learning for Electricity Price Forecasting
A paper presents MetaGrid, a topology-aware transformer that adapts to an unseen…
Context-weighted Discrete Flow Matching
The paper proposes a context-weighted sampler and scaled cross-entropy loss func…