Anonymization: The imperfect science of using data while preserving privacy
Information about us, our actions, and our preferences is created at scale through surveys or
scientific studies or as a result of our interaction with digital devices such as smartphones …
scientific studies or as a result of our interaction with digital devices such as smartphones …
[HTML][HTML] A survey on membership inference attacks and defenses in Machine Learning
Membership inference (MI) attacks mainly aim to infer whether a data record was used to
train a target model or not. Due to the serious privacy risks, MI attacks have been attracting a …
train a target model or not. Due to the serious privacy risks, MI attacks have been attracting a …
Forget to flourish: Leveraging machine-unlearning on pretrained language models for privacy leakage
Fine-tuning large language models on private data for downstream applications poses
significant privacy risks in potentially exposing sensitive information. Several popular …
significant privacy risks in potentially exposing sensitive information. Several popular …
Towards principled assessment of tabular data synthesis algorithms
Data synthesis has been advocated as an important approach for utilizing data while
protecting data privacy. A large number of tabular data synthesis algorithms (which we call …
protecting data privacy. A large number of tabular data synthesis algorithms (which we call …
SoK: Reducing the Vulnerability of Fine-tuned Language Models to Membership Inference Attacks
Natural language processing models have experienced a significant upsurge in recent
years, with numerous applications being built upon them. Many of these applications require …
years, with numerous applications being built upon them. Many of these applications require …
Efficient and Private: Memorisation under differentially private parameter-efficient fine-tuning in language models
O Ma, J Passerat-Palmbach, D Usynin - arXiv preprint arXiv:2411.15831, 2024 - arxiv.org
Fine-tuning large language models (LLMs) for specific tasks introduces privacy risks, as
models may inadvertently memorise and leak sensitive training data. While Differential …
models may inadvertently memorise and leak sensitive training data. While Differential …
SoK: A Systems Perspective on Compound AI Threats and Countermeasures
Large language models (LLMs) used across enterprises often use proprietary models and
operate on sensitive inputs and data. The wide range of attack vectors identified in prior …
operate on sensitive inputs and data. The wide range of attack vectors identified in prior …
[PDF][PDF] Membership Inference Attacks Against Indoor Location Models
V Moghtadaiee, A Fathalizadeh… - Proc. 21st Int. Conf. Secur …, 2024 - scitepress.org
With the widespread adoption of location-based services and the increasing demand for
indoor positioning systems, the need to protect indoor location privacy has become crucial …
indoor positioning systems, the need to protect indoor location privacy has become crucial …
Optimizing Label-Only Membership Inference Attacks by Global Relative Decision Boundary Distances
J Xu, J Hu, C Yu, C Tan - International Conference on Information Security, 2024 - Springer
A sample's distance from the decision boundary is a crucial indicator for predicting whether
a given sample is a member of the training set in label-only membership inference attacks …
a given sample is a member of the training set in label-only membership inference attacks …
Investigating the Effect of Misalignment on Membership Privacy in the White-box Setting
AM Cretu, D Jones, YA de Montjoye, S Tople - arXiv preprint arXiv …, 2023 - arxiv.org
Machine learning models have been shown to leak sensitive information about their training
datasets. Models are increasingly deployed on devices, raising concerns that white-box …
datasets. Models are increasingly deployed on devices, raising concerns that white-box …