English
Related papers

Related papers: PACZero: PAC-Private Fine-Tuning of Language Model…

200 papers

Fine-tuning large language models (LLMs) for specific tasks introduces privacy risks, as models may inadvertently memorise and leak sensitive training data. While Differential Privacy (DP) offers a solution to mitigate these risks, it…

Machine Learning · Computer Science 2024-11-26 Olivia Ma , Jonathan Passerat-Palmbach , Dmitrii Usynin

Deep learning using large models have achieved great success in a wide range of domains. However, training these models on billions of parameters is very challenging in terms of the training speed, memory cost, and communication efficiency,…

Machine Learning · Computer Science 2023-11-21 Zhiqi Bu , Justin Chiu , Ruixuan Liu , Sheng Zha , George Karypis

Large ASR models can inadvertently leak sensitive information, which can be mitigated by formal privacy measures like differential privacy (DP). However, traditional DP training is computationally expensive, and can hurt model performance.…

Cryptography and Security · Computer Science 2024-10-04 Hongbin Liu , Lun Wang , Om Thakkar , Abhradeep Thakurta , Arun Narayanan

Fine-tuning Large Language Models (LLMs) on sensitive datasets carries a substantial risk of unintended memorization and leakage of Personally Identifiable Information (PII), which can violate privacy regulations and compromise individual…

In this paper we develop state-of-the-art privacy attacks against Large Language Models (LLMs), where an adversary with some access to the model tries to learn something about the underlying training data. Our headline results are new…

Cryptography and Security · Computer Science 2024-07-16 Jeffrey G. Wang , Jason Wang , Marvin Li , Seth Neel

Large language models (LLMs) do not preserve privacy at inference-time. The LLM's outputs can inadvertently reveal information about the model's context, which presents a privacy challenge when the LLM is augmented via tools or databases…

Computation and Language · Computer Science 2026-02-03 Rushil Thareja , Preslav Nakov , Praneeth Vepakomma , Nils Lukas

Fine-tuning Large Language Models (LLMs) with first-order methods like back-propagation is computationally intensive. Zeroth-Order (ZO) optimisation uses function evaluations instead of gradients, reducing memory usage, but suffers from…

Computation and Language · Computer Science 2025-07-24 Alessio Galatolo , Zhenbang Dai , Katie Winkle , Meriem Beloucif

Although language models (LMs) demonstrate exceptional capabilities on various tasks, they are potentially vulnerable to extraction attacks, which represent a significant privacy risk. To mitigate the privacy concerns of LMs, machine…

Computation and Language · Computer Science 2024-06-21 Dohyun Lee , Daniel Rim , Minseok Choi , Jaegul Choo

Language models (LMs) may memorize personally identifiable information (PII) from training data, enabling adversaries to extract it during inference. Existing defense mechanisms such as differential privacy (DP) reduce this leakage, but…

Cryptography and Security · Computer Science 2026-02-27 Anthony Hughes , Vasisht Duddu , N. Asokan , Nikolaos Aletras , Ning Ma

Small language models (SLMs) are increasingly valued for their efficiency and deployability in resource-constrained environments, making them useful for on-device, privacy-sensitive, and edge computing applications. On the other hand,…

Artificial Intelligence · Computer Science 2025-08-05 Roya Arkhmammadova , Hosein Madadi Tamar , M. Emre Gursoy

As the capabilities of pre-trained large language models (LLMs) continue to advance, the "pre-train and fine-tune" paradigm has become increasingly mainstream, leading to the development of various fine-tuning methods. However, the privacy…

Computation and Language · Computer Science 2025-07-02 Jie Hou , Chuxiong Wu , Lannan Luo , Qiang Zeng

Membership Inference Attacks (MIA) aim to infer whether a target data record has been utilized for model training or not. Existing MIAs designed for large language models (LLMs) can be bifurcated into two types: reference-free and…

Computation and Language · Computer Science 2024-11-27 Wenjie Fu , Huandong Wang , Chen Gao , Guanghua Liu , Yong Li , Tao Jiang

The first large-scale deployment of private federated learning uses differentially private counting in the continual release model as a subroutine (Google AI blog titled "Federated Learning with Formal Differential Privacy Guarantees"). In…

Machine Learning · Computer Science 2024-02-06 Monika Henzinger , Jalaj Upadhyay , Sarvagya Upadhyay

Data used to train machine learning (ML) models can be sensitive. Membership inference attacks (MIAs), attempting to determine whether a particular data record was used to train an ML model, risk violating membership privacy. ML model…

Cryptography and Security · Computer Science 2022-09-07 Vasisht Duddu , Sebastian Szyller , N. Asokan

Fine-tuning language models (LMs) has yielded success on diverse downstream tasks, but as LMs grow in size, backpropagation requires a prohibitively large amount of memory. Zeroth-order (ZO) methods can in principle estimate gradients using…

Machine Learning · Computer Science 2024-01-12 Sadhika Malladi , Tianyu Gao , Eshaan Nichani , Alex Damian , Jason D. Lee , Danqi Chen , Sanjeev Arora

Large Language Models (LLMs) have the promise to revolutionize computing broadly, but their complexity and extensive training data also expose significant privacy vulnerabilities. One of the simplest privacy risks associated with LLMs is…

Machine Learning · Computer Science 2024-09-25 Rongting Zhang , Martin Bertran , Aaron Roth

While fine-tuning large language models (LLMs) for specific tasks often yields impressive results, it comes at the cost of memory inefficiency due to back-propagation in gradient-based training. Memory-efficient Zeroth-order (MeZO)…

Machine Learning · Computer Science 2026-02-17 Yong Liu , Zirui Zhu , Chaoyu Gong , Minhao Cheng , Cho-Jui Hsieh , Yang You

With the rise of large language models, service providers offer language models as a service, enabling users to fine-tune customized models via uploaded private datasets. However, this raises concerns about sensitive data leakage. Prior…

Cryptography and Security · Computer Science 2026-01-22 Yi Liu , Weixiang Han , Chengjun Cai , Xingliang Yuan , Cong Wang

In this work we analyze the sample complexity of classification by differentially private algorithms. Differential privacy is a strong and well-studied notion of privacy introduced by Dwork et al. (2006) that ensures that the output of an…

Data Structures and Algorithms · Computer Science 2015-09-15 Vitaly Feldman , David Xiao

Reviewers in peer review are often miscalibrated: they may be strict, lenient, extreme, moderate, etc. A number of algorithms have previously been proposed to calibrate reviews. Such attempts of calibration can however leak sensitive…

Cryptography and Security · Computer Science 2022-01-28 Wenxin Ding , Gautam Kamath , Weina Wang , Nihar B. Shah