Search Results for author:
Found 9 papers, 6 papers with code
Watermarking Training Data of Music Generation Models
We study factors that impact the model's generation behaviour: the watermarking technique, the proportion of watermarked samples in the training set, and the robustness of the watermarking technique against the model's tokenizer.
Free Record-Level Privacy Risk Evaluation Through Artifact-Based Methods
Membership inference attacks (MIAs) are widely used to empirically assess privacy risks in machine learning models, both providing model-level vulnerability metrics and identifying the most vulnerable training samples.
SoK: Membership Inference Attacks on LLMs are Rushing Nowhere (and How to Fix It)
We then quantify distribution shifts present in 6 datasets used in the literature using a model-less bag of word classifier and show that all datasets constructed post-hoc suffer from strong distribution shifts.
Certification for Differentially Private Prediction in Gradient-Based Training
Differential privacy upper-bounds the information leakage of machine learning models, yet providing meaningful privacy guarantees has proven to be challenging in practice.
Mosaic Memory: Fuzzy Duplication in Copyright Traps for Large Language Models
This introduces a previously unexplored confounding factor in post-hoc studies of LLM memorization, and questions the effectiveness of (exact) data deduplication as a privacy protection technique.
Copyright Traps for Large Language Models
We here propose to use copyright traps, the inclusion of fictitious entries in original content, to detect the use of copyrighted materials in LLMs with a focus on models where memorization does not naturally occur.
Defending against Reconstruction Attacks with Rényi Differential Privacy
Reconstruction attacks allow an adversary to regenerate data samples of the training set using access to only a trained model.
Opacus: User-Friendly Differential Privacy Library in PyTorch
We introduce Opacus, a free, open-source PyTorch library for training deep learning models with differential privacy (hosted at opacus. ai).
Antipodes of Label Differential Privacy: PATE and ALIBI
We propose two novel approaches based on, respectively, the Laplace mechanism and the PATE framework, and demonstrate their effectiveness on standard benchmarks.
