Engineering risk-based anonymisation solutions for complex data environments
Bibliographic record
Abstract
Technological advancements have dramatically increased the ability to collect, store and process vast quantities of data. The general applicability and precision of analytical tools in artificial intelligence and machine learning have driven organisations to leverage these advances to process personal data in new and innovative ways. As stewards of personal data, organisations need to keep that data safe and ensure processing is legal and appropriate. Having more data, however, has also led to an increased interest to process personal data for purposes other than why they were originally collected, known as secondary purposes. The reuse of personal data introduces important regulatory challenges, increasing the need to disassociate data used for secondary purposes from personal data, be it to safeguard the data, support a legitimate interest, or anonymise the data. Whereas some academics have focused on specific issues preventing more widespread adoption of this privacy-enhancing technology, others have reframed the discussion around anonymisation as risk management. Combining technology-enabled processes with measures of identifiability provides an opportunity to meet complex business needs while ensuring best practice is adopted in reusing sensitive data. This paper examines these many considerations and demonstrates how risk-based anonymisation can and should be detailed, evidence based and objectively supported through measures of identifiability. The engineering of privacy solutions, through the application of risk-based anonymisation, is also briefly explored for complex use cases involving data lakes and hub and spoke data collection, to provide the reader with a deeper understanding of real-world riskbased anonymisation in practice.
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 | 0.018 | 0.044 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.003 | 0.005 |
| Scholarly communication | 0.007 | 0.011 |
| Open science | 0.003 | 0.012 |
| Research integrity | 0.004 | 0.004 |
| 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".