arXiv cs.AI· Abhishek A. Sabnis, Mihai Mitrea, Lya Lugon, Karine Sartelet, Marc Bocquet, Xiaoyuan Cheng, Shupeng Zhu, Sibo Cheng·· 13 小时前AI 评分31
从确定性到生成式深度学习：基于稀疏观测的城市空气质量重建
From Deterministic to Generative Deep Learning for Urban Air Quality Reconstruction from Sparse Observations
AI 导读
研究团队用扩散模型生成框架从稀疏观测重建 NO2、O3、PM2.5 和 PM10 四种污染物的全场分布，并与确定性深度学习模型对比。模型在全场模拟数据上训练，在巴黎 9 至 28 个监测站的真实观测上评估，模拟验证数据上取得高结构相似度，真实观测上生成的空间模式经功率谱分析被认为合理。团队还提出数据增强方法，无需重新训练即可迁移到真实观测并泛化到训练期之外。
正文 · 原文
Abstract:Full-field reconstruction of air pollution is essential for evaluating pollution exposure and supporting public health decision-making. However, the complex interactions among pollutants, hard-to-predict weather patterns, and limited monitoring station coverage make this a complex task. We apply deep learning techniques to provide fast and accurate reconstructions from sparse observations of four key pollutants: NO2, O3, PM2.5 and PM10. Models are trained on full-field simulation data and evaluated on real-world observations collected from 9 to 28 monitoring stations in the city of Paris. We introduce a diffusion-based generative framework for multi-pollutant reconstruction and benchmark its performance against deterministic deep learning models. Despite noisy observations and strong spatial variability, the models achieve high structural similarity on simulated validation data and produce realistic spatial patterns on real-world observations, as indicated by power-spectrum analysis. We introduce data augmentation methods that enable transfer to real-world observations without retraining, allowing the models to generalise beyond the training period. These findings highlight the potential of ML models for reliable real-world deployment in air pollution reconstruction tasks.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.25687 [cs.LG] |
| (or arXiv:2607.25687v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2607.25687 arXiv-issued DOI via DataCite |
Submission history
From: Sibo Cheng [view email]
[v1]
Tue, 28 Jul 2026 13:03:09 UTC (9,388 KB)
[v2]
Mon, 26 Oct 2026 21:54:19 UTC (12,289 KB)
来源：arXiv cs.AI · arxiv.org