Differential Privacy in Tripartite Interaction: A Case Study with Linguistic Minorities in Canada
Bibliographic record
Abstract
Abstract. This paper relates our venture to solve a real-world problem about official language minorities in Canada. The goal was to enable a form of linkage between health data (hosted at ICES – a provincial agency) and language data from the 2006 census (hosted at Statistics Canada – a federal agency) despite a seemingly impossible set of legal constraints. The long-term goal for health researchers is to understand health data according to the linguistic variable, shown to be a health determinant. We first suggested a pattern of tripartite interaction that, by design, prevents collection of residual information by a potential adversary. The suggestion was quickly set aside by Statistics Canada based on the risk of collusion an adversary could exploit among these entities. Our second suggestion was more involved; it consisted in adapting differential privacy mechanisms to the tripartite scheme so as to control the level of leakage in case of collusion. While not being rejected and even receiving enthousiastic interest per se, the solution was considered an option only if other simpler (but also less promising) alternatives are first, and methodically ruled out. 1
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 | 0.007 | 0.017 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.004 |
| Science and technology studies | 0.026 | 0.008 |
| Scholarly communication | 0.005 | 0.002 |
| Open science | 0.002 | 0.005 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.003 | 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 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".