Noise as Signal in Learning from Rare Events
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
Firms increasingly have access to information about the failure events of other firms through public repositories. We study one such repository that accumulates reports of adverse events in the medical device industry. We provide qualitative evidence that shows how firms select a sample of adverse events and then engage in inferential learning. We show that firms use the reports of others to extract new valid knowledge from the adverse events in other firms. We use quantitative evidence to explore how a public repository can be used to provide more direct evidence of vicarious learning. Our findings challenge some standard assumptions about vicarious learning. First, we show that the learning in a repository does not come from referent others. Instead, it emerges directly from failure events that might ordinarily be dismissed as noise. Second, we show that the learning does not come from copying others. Instead, it is constructed by firm members as they assemble individual failure events to identify possibilities they had not considered. Third, in contrast to vicarious learning, where the referent others and rare events provide the context, repository-based learning requires that actors impose their own context as part of the learning process. Our qualitative and quantitative evidence serve explanatory purposes by showing how firms use a repository of failure events to identify moments of valid learning, and they serve exploratory purposes by investigating how we can demonstrate reliable learning from a repository of failure events. The e-companion is available at https://doi.org/10.1287/orsc.2017.1179 .
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 | 0.005 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.005 |
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