MétaCan
Menu
Back to cohort
Record W4401164363 · doi:10.1109/csp62567.2024.00010

Linkage Deanonymization Risks, Data-Matching and Privacy: A Case Study

2024· article· en· W4401164363 on OpenAlex

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.

affAt least one author lists a Canadian institution in the pinned OpenAlex snapshot.
aboutThe title or abstract carries a Canadian signal from the geographic lexicon.

Bibliographic record

Venuenot available
Typearticle
Languageen
FieldSocial Sciences
TopicNames, Identity, and Discrimination Research
Canadian institutionsNew York Institute of Technology
Fundersnot available
KeywordsComputer scienceLinkage (software)Record linkageMatching (statistics)Information privacyComputer securityInternet privacyStatisticsEnvironmental healthMathematicsGeneticsMedicine

Abstract

fetched live from OpenAlex

Data anonymization is a widely used solution for preserving users' privacy in data publishing. It is achieved through Privacy Preserving Data Publishing (PPDP), which provides a set of models, tools, and methods to defend against the privacy threats. However, data anonymization has been considered insufficient to protect data privacy for years. New de-anonymization techniques increase the risk of breaching privacy protection through data anonymization. Previous researchers write papers on de-anonymization techniques. In most cases, the dataset used for simulation had a limited range and small scale. In this paper, we aim to evaluate the potential of large-scale datasets being vulnerable to linkage attacks with external social media network data. The data set resource comes from the April 2023 Labour Force Survey (Statistics Canada, 2023). The team estimates the probability of successful linkage attacks of a dataset drawn from the entire population with published attributes that could be linked with publicly available data with the same individually specific attributes. The result shows that the estimated deanonymization risks for each of Newfoundland and Labrador, Prince Edward Island, and Manitoba are 12.5%, 12.5% and 1.9% respectively. It can be summarized that de-anonymization risk is inversely proportional to the underlying population size (or number of entries). It is intuitive in that the more diverse the population, the less risk of deanonymization through linkage with selected attributes. Therefore, we conclude that a more extensive scaling and more comprehensive entries dataset decreases the probability of successful linkage de-anonymization attacks.

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 imitation

Not 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.

metaresearch head score (Codex)0.002
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Qualitative · Consensus signal: Qualitative
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.449
Threshold uncertainty score0.969

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0020.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.000
Science and technology studies0.0010.000
Scholarly communication0.0010.001
Open science0.0000.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

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.

Opus teacher head0.225
GPT teacher head0.498
Teacher spread0.273 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it