Response to comment on “Do we need flexible machine-learning algorithms to assess the effect of long-term exposure to fine particulate matter on mortality?”
Bibliographic record
Abstract
We thank Williams et al1 for their attention to our manuscript and for pointing out some technical concerns in our manuscript. Regarding their first point, we agree with the definitions of bias and variance that they propose as expectations over many iterations, which therefore cannot be observed directly in a single study. However, as stated in our manuscript, we sought to report observed errors for a few estimators in a single real-world case study. We were unable to detect any meaningful advantage of using a flexible machine-learning algorithm in this one instance, considering a range of (unknown) true values for the causal effect parameter. We agree that this is just one instance and that one cannot easily generalize to other settings. Their recommendation to pursue this question in a plasmode simulation by sampling from the real dataset is an exciting idea that we hope to pursue. Regarding the second point, we also agree that methodological developments are still needed regarding non–smooth statistical estimands. Environmental regulations are often expressed as thresholds, making the available approaches inapt for the causal questions that most closely relate to what air pollution epidemiologists need to know. We hope our manuscript and the letter by Williams et al1 can help to motivate further developments in flexible modeling that will more directly align with real-world policy questions in environmental epidemiology. Conflicts of interest statement The authors declare that they have no conflicts of interest with regard to the content of this report.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.059 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.003 | 0.003 |
| Scholarly communication | 0.003 | 0.004 |
| Open science | 0.004 | 0.002 |
| Research integrity | 0.038 | 0.046 |
| Insufficient payload (model declined to judge) | 0.013 | 0.014 |
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 source (direct Gemma or distilled Codex), 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".