Secure Approximate String Matching for Privacy-Preserving Record Linkage
Bibliographic record
Abstract
Real-world applications of record linkage often require matching to be robust in spite of small variations in string fields. For example, two health care providers should be able to detect a patient in common, even if one record contains a typo or transcription error. In the privacy-preserving setting, however, the problem of approximate string matching has been cast as a trade-off between security and practicality, and the literature has mainly focused on Bloom filter encodings, an approach which can leak significant information about the underlying records. We present a novel public-key construction for secure two-party evaluation of threshold functions in restricted domains based on embeddings found in the message spaces of additively homomorphic encryption schemes. We use this to construct an efficient two-party protocol for privately computing the threshold Dice coefficient. Relative to the approach of Bloom filter encodings, our proposal offers formal security guarantees and greater matching accuracy. We implement the protocol and demonstrate the feasibility of this approach in linking medium-sized patient databases with tens of thousands of records.
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 | 0.023 |
| Meta-epidemiology (narrow) | 0.000 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.003 |
| Science and technology studies | 0.002 | 0.003 |
| Scholarly communication | 0.004 | 0.009 |
| Open science | 0.002 | 0.008 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".