Privacy Preserving Machine Learning using Deidentification Techniques for HR and Payroll Data
Bibliographic record
Abstract
The use of machine learning in Human Resources (HR) and payroll systems represents significant capability gains for automation, decision-making, and operational efficiencies.However, this transition process raises significant concerns related to the protection of sensitive employee data.Privacy-preserving computer scientific methods, like deidentification (removing or altering identifiers such that the individual cannot be identified), are becoming an increasingly important way to preserve privacy while maintaining the richness of datasets for use in analytical research.The purpose of this review article is to provide a comprehensive report on various privacy-preserving techniques, especially de-identification techniques common to HR and payroll data.The review will include real-world examples of privacy-preserving methods, including CV de-identification, clustering of quasi-identifiers (QI), narrative-level deidentification, and federated learning (model training without the data leaving the local source, but sharing model updates).The article looks at implications for privacy and utility trade-offs, assesses potential reidentification risks, and summarizes innovations in secured data, such as format-preserving transformations (anonymizing and preserving values such as IDs/dates) and voice anonymization.The review article considers issues identified in the current literature, as well as policy implications and future directions for utilizing secure and compliant machine learning frameworks in human resource settings.
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 | 0.011 | 0.028 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.005 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.005 | 0.007 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.002 | 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".