Accurate and Scalable Gaussian Processes for Fine-grained Air Quality Inference

Download paper here

Air pollution is a global problem and has a severe impact on human health. Fine-grained air quality (AQ) monitoring is important in mitigating air pollution. However, existing AQ station deployments are sparse. Conventional interpolation techniques fail to learn the complex AQ phenomena. Physics-based models require domain knowledge and pollution source data for AQ modeling. In this work, we propose a Gaussian processes based approach for estimating AQ. The important features of our approach are: a) a non-stationary (NS) kernel to allow input depended smoothness of fit; b) a Hamming distance-based kernel for categorical features; and c) a locally periodic kernel to capture temporal periodicity. We leverage batch-wise training to scale our approach to a large amount of data. Our approach outperforms the conventional baselines as well as a state-of-the-art neural attention-based approach.

Recommended citation: Zeel Patel, Harsh Patel*, Palak Purohit*, Shivam Sahni*, Nipun Batra. Accurate and Scalable Gaussian Processes for Fine-grained Air Quality Inference." Thirty-Sixth Association for the Advancement of Artificial Intelligence (AAAI) Conference [AAAI 2022].
Download Paper