A New Method of Estimating Global PM2.5 Concentrations using Satellite Images
Bibliographic record
Abstract
OPS 49: Air pollution exposure methods, Room 412, Floor 4, August 27, 2019, 10:30 AM - 12:00 PM Background: Few studies have examined deep learning image analysis for use in evaluating environmental exposures Methods: Here we present a new method of predicting spatial variations in outdoor fine particulate air pollution (PM2.5) concentrations using deep convolutional neural networks. Specifically, we trained new deep learning models over the global exposure range (<1-436 ug/m3) using a large database of satellite images paired with ground level PM2.5 measurements available from the World Health Organization. In addition, we trained models over the more limited exposure range of North America (<1-16 ug/m3) using a large database of satellite images linked to ground-level PM2.5 estimates obtained through remote sensing. Final model selection was based on a systematic evaluation of well-known architectures for the convolutional base including InceptionV3, Xception, and VGG16. Models were developed to predict continuous exposures as well as categorial estimates across ten ordinal categories of exposure split by deciles. Results: The Xception architecture performed best in both models. For the global model, a root mean square error (RMSE) value of 13.01 ug/m3 was observed with a strong correlation between measured and predicted values in the disjoint test set (R2=0.75). For North America, the RMSE value was 0.74 ug/m3 with an R2 value of 0.89. Conclusions: Our findings suggest that deep convolutional neural networks may offer an alternative, cost-effective means of predicting spatial variations in long-term average PM2.5 concentrations.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".