Beyond Theoretical Bounds: Empirical Privacy Loss Calibration for Text Rewriting Under Local Differential Privacy
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915886802141184 |
|---|---|
| author | Li, Weijun Sébert, Arnaud Grivet Xu, Qiongkai McIver, Annabelle Dras, Mark |
| author_facet | Li, Weijun Sébert, Arnaud Grivet Xu, Qiongkai McIver, Annabelle Dras, Mark |
| contents | The growing use of large language models has increased interest in sharing textual data in a privacy-preserving manner. One prominent line of work addresses this challenge through text rewriting under Local Differential Privacy (LDP), where input texts are locally obfuscated before release with formal privacy guarantees. These guarantees are typically expressed by a parameter $\varepsilon$ that upper bounds the worst-case privacy loss. However, nominal $\varepsilon$ values are often difficult to interpret and compare across mechanisms. In this work, we investigate how to empirically calibrate across text rewriting mechanisms under LDP. We propose TeDA, which formulates calibration via a hypothesis-testing framework that instantiates text distinguishability audits in both surface and embedding spaces, enabling empirical assessment of indistinguishability from privatized texts. Applying this calibration to several representative mechanisms, we demonstrate that similar nominal $\varepsilon$ bounds can imply very different levels of distinguishability. Empirical calibration thus provides a more comparable footing for evaluating privacy-utility trade-offs, as well as a practical tool for mechanism comparison and analysis in real-world LDP text rewriting deployments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_22968 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Beyond Theoretical Bounds: Empirical Privacy Loss Calibration for Text Rewriting Under Local Differential Privacy Li, Weijun Sébert, Arnaud Grivet Xu, Qiongkai McIver, Annabelle Dras, Mark Cryptography and Security Computation and Language The growing use of large language models has increased interest in sharing textual data in a privacy-preserving manner. One prominent line of work addresses this challenge through text rewriting under Local Differential Privacy (LDP), where input texts are locally obfuscated before release with formal privacy guarantees. These guarantees are typically expressed by a parameter $\varepsilon$ that upper bounds the worst-case privacy loss. However, nominal $\varepsilon$ values are often difficult to interpret and compare across mechanisms. In this work, we investigate how to empirically calibrate across text rewriting mechanisms under LDP. We propose TeDA, which formulates calibration via a hypothesis-testing framework that instantiates text distinguishability audits in both surface and embedding spaces, enabling empirical assessment of indistinguishability from privatized texts. Applying this calibration to several representative mechanisms, we demonstrate that similar nominal $\varepsilon$ bounds can imply very different levels of distinguishability. Empirical calibration thus provides a more comparable footing for evaluating privacy-utility trade-offs, as well as a practical tool for mechanism comparison and analysis in real-world LDP text rewriting deployments. |
| title | Beyond Theoretical Bounds: Empirical Privacy Loss Calibration for Text Rewriting Under Local Differential Privacy |
| topic | Cryptography and Security Computation and Language |
| url | https://arxiv.org/abs/2603.22968 |