Learnijoy
ResearchAI ResearchAI Engineering & DevTools

New Dataset Boosts Global Air Quality Forecasting with AI Models

Rishi Bharadwaj, Manik Gupta, Pandarasamy Arjunan· July 23, 2026 View original

Summary

Researchers introduce Air Quality Arena (AQA), a large-scale, multi-country, multi-pollutant dataset and benchmark for air quality forecasting. AQA enables evaluation of time series foundation models (TSFMs) on real-world data, demonstrating their superior performance over classical baselines.

Air pollution poses a significant global health threat, making accurate forecasting crucial. While machine learning is increasingly applied to this challenge, existing benchmarks for air quality prediction often lack geographic and pollutant diversity, and fail to assess the latest time series foundation models (TSFMs) on large-scale, real-world data. To address these gaps, a new initiative called Air Quality Arena (AQA) has been launched. AQA comprises a comprehensive dataset (AQA-Data) covering six major pollutants over three years across seven diverse countries and four continents, encompassing over 14,000 station-pollutant series. It also includes a benchmark (AQA-Bench) for evaluating short-term air quality forecasting. Initial evaluations on AQA demonstrate that TSFMs are highly effective zero-shot forecasters, consistently outperforming traditional baselines. Notably, the top-performing model utilizes a cross-modal architecture, integrating a vision foundation model for time series forecasting. This public release aims to provide a robust resource for advancing air quality prediction research.

Why it matters

Improved air quality forecasting directly impacts public health, urban planning, and environmental policy, offering professionals better tools for risk assessment and mitigation strategies.

How to implement this in your domain

  1. 1Utilize the AQA dataset to train and benchmark new air quality forecasting models.
  2. 2Explore cross-modal architectures, integrating vision models with time series data for enhanced predictions.
  3. 3Collaborate with environmental agencies to deploy advanced forecasting models for public health initiatives.
  4. 4Develop real-time dashboards and alert systems based on improved air quality predictions.

Who benefits

Environmental MonitoringPublic HealthUrban PlanningAI DevelopmentGovernment

Key takeaways

  • A new large-scale dataset, Air Quality Arena (AQA), is available for air quality forecasting.
  • Time Series Foundation Models (TSFMs) significantly outperform classical baselines in this domain.
  • Cross-modal architectures, including vision models, show promise for advanced time series forecasting.
  • Accurate air quality forecasting is vital for public health and environmental management.

Original post by Rishi Bharadwaj, Manik Gupta, Pandarasamy Arjunan

"arXiv:2607.19381v1 Announce Type: new Abstract: Air pollution causes an estimated 7.9 million premature deaths annually, making accurate forecasting a critical public health priority. Machine learning is increasingly being applied to forecast air pollution levels, yet existing be…"

View on X

Originally posted by Rishi Bharadwaj, Manik Gupta, Pandarasamy Arjunan on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses

More in AI Research

AI ResearchAI Engineering & DevTools

New Adaptive Filter Improves Time-Series Prediction with Input Noise

Researchers developed the RFFBCGA algorithm, a new nonlinear adaptive filter that effectively mitigates both input and output noise in time-series prediction. This method maintains a fixed network structure while enhancing robustness across various noise scenarios.

Yi Peng, Haiquan Zhao, Jinhui HuJul 23, 2026
AI ResearchAI Engineering & DevTools

New Algorithm Learns Local Causal Structures with Latent Variables

Researchers propose LoCaLS, a new algorithm for learning local causal structures around a target variable from observational data, even when latent variables and selection bias are present. LoCaLS achieves high accuracy with significantly less computational effort than global causal discovery methods.

Zheng Li, Hao Zhang, Ruxin Wang, Ruichu Cai, Kun Zhang, Feng XieJul 23, 2026
AI Engineering & DevToolsAI Research

New Framework Evaluates AI Robustness with Minimum-Norm Attacks

Researchers introduce a unified framework for evaluating adversarial robustness using a comprehensive pool of minimum-norm attacks and robustness-perturbation curves across multiple norms. This approach addresses limitations of fixed-epsilon evaluations, providing a more stable and controllable assessment of AI model defenses.

Luca Scionis, Luca Melis, Maura Pintor, Fabio Brau, Ambra Demontis, Giorgio Fumera, Fabio Roli, Battista BiggioJul 23, 2026