Learning air pollution with bidirectional LSTM RNN

An accurate understanding of air pollutants in a continuous space-time domain by spatiotemporal interpolation is critical for meaningful assessment of the quantitative relationship between the public health and perennial environmental exposures. Existing spatiotemporal interpolation algorithms are usually based on unrealistic assumptions by restricting the interpolation models to the ones with explicit and simple mathematical descriptions, thus neglecting plenty of hidden yet critical influence factors. We developed an efficient deep-learning-based spatiotemporal interpolation algorithm which can generate more accurate estimation for air pollution on a large geographic scale and over a long time period. The experimental results demonstrate the efficacy and efficiency of our novel algorithm.