MDM-Privacy: A Derivative-Driven Framework for Differential Privacy
Bibliographic record
Abstract
Ensuring rigorous privacy protection while maintaining data utility remains a central challenge in privacy-preserving data analysis. Traditional Differential Privacy mechanisms, such as the Laplace and Gaussian mechanisms, rely on additive noise calibrated to sensitivity, which often leads to degraded utility, particularly in high-dimensional or low ε regimes. Moreover, these mechanisms lack the flexibility to adapt to the local structure or the gradient behavior of the data. This paper introduces the MDM-Privacy framework, which leverages the ε-MDM-Privacy mechanism, to calibrate privacy noise following sample-level sensitivity. Unlike traditional additive noise mechanisms, ε-MDM-Privacy employs a non-additive, exponential noise model where the privatized output is computed as a function of the formM(x) =C∙eε∙Lp(x), enabling precise control over the privacy-utility trade-off through both the privacy budget ε and the norm orderp. The constantCis estimated via Monte Carlo simulations to ensure probabilistic utility guarantees under confidence bounds. Extensive experiments on the UCI Adult dataset demonstrate that ε-MDM-Privacy achieves significantly lower Mean Absolute Error (MAE) and Mean Relative Error (MRE) compared to the classical Laplace mechanism, particularly under tight privacy regimes. The MDM-Privacy also exhibits robustness across different Minkowski norms, confirming its adaptability to various sensitivity geometries. Moreover, the ε-MDM-Privacy is consistent under both parallel and sequential compositions, enabling the execution of an unlimited number of queries without exhausting the privacy budget (ε). These results highlight the effectiveness of ε-MDM-Privacy as a tunable and utility-preserving alternative for differentially private data release.
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 | 0.019 | 0.038 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.002 | 0.003 |
| Science and technology studies | 0.002 | 0.005 |
| Scholarly communication | 0.006 | 0.009 |
| Open science | 0.005 | 0.009 |
| Research integrity | 0.003 | 0.007 |
| Insufficient payload (model declined to judge) | 0.004 | 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".