Mitigating Authorship Attribution in Social Media Texts: A Privacy by Translation Approach
Bibliographic record
Abstract
The exponential growth of social media usage has raised significant privacy concerns, especially regarding adversarial authorship attribution. Social media texts pose distinctive privacy risks for authorship attribution due to their informal, high-variance style. We study Privacy by Translation (PbT) for this setting and its evaluated hybrid with Differential Privacy (PbT+DP) to mitigate authorship identification while preserving utility for social media mental-health classification. Our method transforms social media texts into semantically equivalent forms that obscure authorial style. PbT is formulated in a two-step process (translation and learning) to optimize the trade-off between privacy and utility. We assess the effectiveness of our method across various languages and compare it with existing methods. Our findings reveal that a hybrid approach combining DP with PbT outperforms standalone PbT, DP, and existing methods in balancing privacy and utility in social media mental health classification. The hybrid method reduces adversarial authorship classification by 89.1% and maintains an efficient privacy-utility ratio of 0.40. We also investigate the transformation of linguistic features, including stylistic, syntactic, and semantic elements, to understand their roles in authorship attribution.
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 | 0.010 | 0.049 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.002 | 0.003 |
| Scholarly communication | 0.004 | 0.006 |
| Open science | 0.002 | 0.004 |
| 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".