Abstract
Large pre-trained language models dominate the current state-of-the-art for many natural language processing applications, including the field of clinical NLP. Several studies have found that these can be susceptible to privacy attacks that are unacceptable in the clinical domain where personally identifiable information (PII) must not be exposed. However, there is no consensus regarding how to quantify the privacy risks of different models. One prominent suggestion is to quantify these risks using membership inference attacks. In this study, we show that a state-of-the-art membership inference attack on a clinical BERT model fails to detect the privacy benefits from pseudonymizing data. This suggests that such attacks may be inadequate for evaluating token-level privacy preservation of PIIs.- Anthology ID:
- 2023.nodalida-1.33
- Volume:
- Proceedings of the 24th Nordic Conference on Computational Linguistics (NoDaLiDa)
- Month:
- May
- Year:
- 2023
- Address:
- Tórshavn, Faroe Islands
- Editors:
- Tanel Alumäe, Mark Fishel
- Venue:
- NoDaLiDa
- SIG:
- Publisher:
- University of Tartu Library
- Note:
- Pages:
- 318–323
- Language:
- URL:
- https://aclanthology.org/2023.nodalida-1.33
- DOI:
- Cite (ACL):
- Thomas Vakili and Hercules Dalianis. 2023. Using Membership Inference Attacks to Evaluate Privacy-Preserving Language Modeling Fails for Pseudonymizing Data. In Proceedings of the 24th Nordic Conference on Computational Linguistics (NoDaLiDa), pages 318–323, Tórshavn, Faroe Islands. University of Tartu Library.
- Cite (Informal):
- Using Membership Inference Attacks to Evaluate Privacy-Preserving Language Modeling Fails for Pseudonymizing Data (Vakili & Dalianis, NoDaLiDa 2023)
- PDF:
- https://preview.aclanthology.org/ingest-acl-2023-videos/2023.nodalida-1.33.pdf