English
Related papers

Related papers: DP-InstaHide: Provably Defusing Poisoning and Back…

200 papers

Data poisoning is one of the most relevant security threats against machine learning and data-driven technologies. Since many applications rely on untrusted training data, an attacker can easily craft malicious samples and inject them into…

Cryptography and Security · Computer Science 2021-12-01 Nicolas M. Müller , Simon Roschmann , Konstantin Böttinger

Data poisoning and backdoor attacks manipulate training data in order to cause models to fail during inference. A recent survey of industry practitioners found that data poisoning is the number one concern among threats ranging from model…

Machine Learning · Computer Science 2021-06-18 Avi Schwarzschild , Micah Goldblum , Arjun Gupta , John P Dickerson , Tom Goldstein

Knowledge distillation has become a cornerstone in modern machine learning systems, celebrated for its ability to transfer knowledge from a large, complex teacher model to a more efficient student model. Traditionally, this process is…

Cryptography and Security · Computer Science 2026-01-13 Chen Wu , Qian Ma , Prasenjit Mitra , Sencun Zhu

Recent deep neural networks (DNNs) have came to rely on vast amounts of training data, providing an opportunity for malicious attackers to exploit and contaminate the data to carry out backdoor attacks. However, existing backdoor attack…

Cryptography and Security · Computer Science 2024-04-22 Ziqiang Li , Hong Sun , Pengfei Xia , Heng Li , Beihao Xia , Yi Wu , Bin Li

Prior work on differential privacy analysis of randomized SGD algorithms relies on composition theorems, where the implicit (unrealistic) assumption is that the internal state of the iterative algorithm is revealed to the adversary. As a…

Machine Learning · Statistics 2022-10-18 Jiayuan Ye , Reza Shokri

Fine-tuning large language models on downstream tasks is crucial for realizing their cross-domain potential but often relies on sensitive data, raising privacy concerns. Differential privacy (DP) offers rigorous privacy guarantees and has…

Machine Learning · Computer Science 2026-01-19 Lele Zheng , Xiang Wang , Tao Zhang , Yang Cao , Ke Cheng , Yulong Shen

Although local differential privacy (LDP) protects individual users' data from inference by an untrusted data curator, recent studies show that an attacker can launch a data poisoning attack from the user side to inject carefully-crafted…

Cryptography and Security · Computer Science 2023-03-13 Xiaoguang Li , Ninghui Li , Wenhai Sun , Neil Zhenqiang Gong , Hui Li

Fine-tuning large language models (LLMs) has become an essential strategy for adapting them to specialized tasks; however, this process introduces significant privacy challenges, as sensitive training data may be inadvertently memorized and…

Cryptography and Security · Computer Science 2025-05-02 Hao Du , Shang Liu , Yang Cao

Federated learning is distributed model training across several clients without disclosing raw data. Despite advancements in data privacy, risks still remain. Differential Privacy (DP) is a technique to protect sensitive data by adding…

Machine Learning · Computer Science 2025-10-14 Tejash Varsani

This study investigates privacy leakage in dimensionality reduction methods through a novel machine learning-based reconstruction attack. Employing an informed adversary threat model, we develop a neural network capable of reconstructing…

Cryptography and Security · Computer Science 2025-06-03 Chayadon Lumbut , Donlapark Ponnoprat

Privacy is a growing concern in modern deep-learning systems and applications. Differentially private (DP) training prevents the leakage of sensitive information in the collected training data from the trained machine learning models. DP…

Machine Learning · Computer Science 2024-08-28 Xinwei Zhang , Zhiqi Bu , Mingyi Hong , Meisam Razaviyayn

As more and more artificial intelligence (AI) technologies move from the laboratory to real-world applications, the open-set and robustness challenges brought by data from the real world have received increasing attention. Data augmentation…

Machine Learning · Computer Science 2022-12-09 Zhendong Liu , Wenyu Jiang , Min guo , Chongjun Wang

Model inversion attacks pose a significant privacy threat to machine learning models by reconstructing sensitive data from their outputs. While various defenses have been proposed to counteract these attacks, they often come at the cost of…

Cryptography and Security · Computer Science 2024-12-11 Shuai Zhou , Dayong Ye , Tianqing Zhu , Wanlei Zhou

Backdoor and data poisoning attacks can achieve high attack success while evading existing spectral and optimisation based defences. We show that this behaviour is not incidental, but arises from a fundamental geometric mechanism in input…

Machine Learning · Statistics 2026-02-03 Diego Granziol

Backdoor attacks pose serious security threats to deep neural networks (DNNs). Backdoored models make arbitrarily (targeted) incorrect predictions on inputs embedded with well-designed triggers while behaving normally on clean inputs. Many…

Cryptography and Security · Computer Science 2023-07-21 Yudong Gao , Honglong Chen , Peng Sun , Junjian Li , Anqing Zhang , Zhibo Wang

Differential Privacy (DP) provides an elegant mathematical framework for defining a provable disclosure risk in the presence of arbitrary adversaries; it guarantees that whether an individual is in a database or not, the results of a DP…

Cryptography and Security · Computer Science 2021-08-19 Aleksandra Slavkovic , Roberto Molinari

Differential privacy (DP) has become the standard for private data analysis. Certain machine learning applications only require privacy protection for specific protected attributes. Using naive variants of differential privacy in such use…

Cryptography and Security · Computer Science 2025-06-25 Saeed Mahloujifar , Chuan Guo , G. Edward Suh , Kamalika Chaudhuri

A private machine learning algorithm hides as much as possible about its training data while still preserving accuracy. In this work, we study whether a non-private learning algorithm can be made private by relying on an instance-encoding…

Data augmentation is used extensively to improve model generalisation. However, reliance on external libraries to implement augmentation methods introduces a vulnerability into the machine learning pipeline. It is well known that backdoors…

Machine Learning · Computer Science 2022-10-03 Joseph Rance , Yiren Zhao , Ilia Shumailov , Robert Mullins

Backdoor attacks undermine the integrity of machine learning models by allowing attackers to manipulate predictions using poisoned training data. Such attacks lead to targeted misclassification when specific triggers are present, while the…

Cryptography and Security · Computer Science 2025-02-12 Shaokui Wei , Jiayin Liu , Hongyuan Zha