Ew-tune: A framework for privately fine-tuning large language models with differential privacy

R Behnia, MR Ebrahimi, J Pacheco… - … Conference on Data …, 2022 - ieeexplore.ieee.org
Pre-trained Large Language Models (LLMs) are an integral part of modern AI that have led
to breakthrough performances in complex AI tasks. Major AI companies with expensive …

Differentially private fine-tuning of language models

D Yu, S Naik, A Backurs, S Gopi, HA Inan… - arXiv preprint arXiv …, 2021 - arxiv.org
We give simpler, sparser, and faster algorithms for differentially private fine-tuning of large-
scale pre-trained language models, which achieve the state-of-the-art privacy versus utility …

Large language models can be good privacy protection learners

Y Xiao, Y Jin, Y Bai, Y Wu, X Yang, X Luo, W Yu… - arXiv preprint arXiv …, 2023 - arxiv.org
The proliferation of Large Language Models (LLMs) has driven considerable interest in fine-
tuning them with domain-specific data to create specialized language models. Nevertheless …

Dp-forward: Fine-tuning and inference on language models with differential privacy in forward pass

M Du, X Yue, SSM Chow, T Wang, C Huang… - Proceedings of the 2023 …, 2023 - dl.acm.org
Differentially private stochastic gradient descent (DP-SGD) adds noise to gradients in back-
propagation, safeguarding training data from privacy leakage, particularly membership …

Differentially private decoding in large language models

J Majmudar, C Dupuy, C Peris, S Smaili… - arXiv preprint arXiv …, 2022 - arxiv.org
Recent large-scale natural language processing (NLP) systems use a pre-trained Large
Language Model (LLM) on massive and diverse corpora as a headstart. In practice, the pre …

Privacy-preserving prompt tuning for large language model services

Y Li, Z Tan, Y Liu - arXiv preprint arXiv:2305.06212, 2023 - arxiv.org
Prompt tuning provides an efficient way for users to customize Large Language Models
(LLMs) with their private data in the emerging LLM service scenario. However, the sensitive …

Learning differentially private recurrent language models

HB McMahan, D Ramage, K Talwar… - arXiv preprint arXiv …, 2017 - arxiv.org
We demonstrate that it is possible to train large recurrent language models with user-level
differential privacy guarantees with only a negligible cost in predictive accuracy. Our work …

Differentially private language models benefit from public pre-training

G Kerrigan, D Slack, J Tuyls - arXiv preprint arXiv:2009.05886, 2020 - arxiv.org
Language modeling is a keystone task in natural language processing. When training a
language model on sensitive information, differential privacy (DP) allows us to quantify the …

Natural language understanding with privacy-preserving bert

C Qu, W Kong, L Yang, M Zhang, M Bendersky… - Proceedings of the 30th …, 2021 - dl.acm.org
Privacy preservation remains a key challenge in data mining and Natural Language
Understanding (NLU). Previous research shows that the input text or even text embeddings …

Privacy preserving large language models: Chatgpt case study based vision and framework

I Ullah, N Hassan, SS Gill, B Suleiman… - arXiv preprint arXiv …, 2023 - arxiv.org
The generative Artificial Intelligence (AI) tools based on Large Language Models (LLMs) use
billions of parameters to extensively analyse large datasets and extract critical private …