Predicting population trends of birds worldwide with big data and machine learning
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Birds are crucial for the functioning of Earth’s ecosystems but bird population declines have been documented worldwide in recent decades. A global assessment of potential causes of population declines is needed. Our goal here was to combine the power of big data and machine learning to identify predictors correlated with bird population declines and to predict population declines for species with unknown population trends on the IUCN Red List. From existing online databases, we gathered detailed species‐level data for 10 964 extant bird species around the world, featuring life history, ecology, distribution, taxonomy and categorical population trend information (i.e. decreasing or not decreasing). For the 10 163 species with known population trends, we split the data into a 75% training set to tune and train a machine‐learning model (Light Gradient Boosting Machine – ‘LightGBM’) and a 25% test set to evaluate the trained model. Our model predicted (i) bird population declines with an ROC AUC score of 0.828, F1 score of 0.748 and average accuracy of 0.747, and (ii) that 47% ( n = 801) of bird species with currently unknown population trends are declining. Correlation analyses suggested that, globally, the top predictor associated with bird population declines was a severely fragmented population, with non‐migratory birds in South American and Southeast Asian tropical and subtropical forests being particularly vulnerable. Despite the lack of long‐term quantitative population trend data for all species worldwide, our study presents big data and machine learning as a useful tool for informing conservation priorities, lending insight, albeit imperfect, into bird population declines on the global scale for the first time.
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.
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 | 0.000 | 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) | 0.014 | 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 it