Evaluation Robust but Robust to What?
Bibliographic record
Abstract
Generalization is at the core of evaluation, we estimate the performance of a model on data we have never seen but expect to encounter later on. Our current evaluation procedures assume that the data already seen is a random sample of the domain from which all future data will be drawn. Unfortunately, in practical situations this is rarely the case. Changes in the underlying probabilities will occur and we must evaluate how robust our models to such differences. This paper takes the position that models should be robust in two senses. Firstly, that any small changes in the joint probabilities should not cause large changes in performance. Secondly, that when the dependencies between attributes and the class are constant and only the marginal change, simple adjustments should be sufficient to restore a model's performance. This paper is intended to generate debate on how measures of robustness might become part of our normal evaluation procedures. Certainly some clear demonstrations of robustness would improve our confidence in our models' practical merits.
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 | 0.127 | 0.375 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.002 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.006 |
| Scholarly communication | 0.011 | 0.016 |
| Open science | 0.003 | 0.005 |
| Research integrity | 0.004 | 0.005 |
| Insufficient payload (model declined to judge) | 0.004 | 0.002 |
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".