Attack Algorithm

Attack algorithms target the vulnerabilities of machine learning models, aiming to manipulate their outputs or steal data. Current research focuses on developing increasingly sophisticated attacks against various model types, including large language models, federated learning systems, and deep neural networks used in image recognition and other applications, employing techniques like gradient manipulation, universal perturbations, and data poisoning. These advancements highlight the critical need for robust defenses and underscore the significant implications for the security and reliability of AI systems across diverse sectors.

Papers

April 30, 2024

Leveraging Label Information for Stealthy Data Stealing in Vertical Federated Learning
Duanyi Yao, Songze Li, Xueluan Gong, Sizai Hou, Gaoning Pan
Vertical Federated Learning Label Information Bi Discriminator Auxiliary Classifier Attack Algorithm Malicious Attack

February 24, 2024

PRP: Propagating Universal Perturbations to Attack Large Language Model Guard-Rails
Neal Mangaokar, Ashish Hooda, Jihye Choi, Shreyas Chandrashekaran, Kassem Fawaz, Somesh Jha, Atul Prakash
Large Language Model Threat Model Universal Adversarial Attack Algorithm Universal Perturbation

February 5, 2024

Time-Distributed Backdoor Attacks on Federated Spiking Learning
Gorka Abad, Stjepan Picek, Aitor Urbieta
Backdoor Attack Attack Performance Attack Algorithm

February 2, 2024

Preference Poisoning Attacks on Reward Model Learning
Junlin Wu, Jiongxiao Wang, Chaowei Xiao, Chenguang Wang, Ning Zhang, Yevgeniy Vorobeychik
Reward Model Preference Learning Pairwise Comparison Preference Rating Attack Algorithm Preference Poisoning

September 18, 2023

Spoofing attack augmentation: can differently-trained attack models improve generalisation?
Wanying Ge, Xin Wang, Junichi Yamagishi, Massimiliano Todisco, Nicholas Evans
Strong Generalization Deepfake Detection Attack Algorithm Enhancement Attack

June 5, 2023

Adversarial Ink: Componentwise Backward Error Attacks on Deep Learning
Lucas Beerens, Desmond J. Higham
Deep Learning Deep Neural Network Adversarial Attack Adversarial Learning Attack Algorithm

March 22, 2023

Reliable and Efficient Evaluation of Adversarial Robustness for Deep Hashing-Based Retrieval
Xunguang Wang, Jiawang Bai, Xinyue Xu, Xiaomeng Li
Adversarial Robustness Efficient Evaluation Deep Hashing Attack Algorithm Permutation Hashing

January 31, 2023

Reverse engineering adversarial attacks with fingerprints from adversarial examples
David Aaron Nicholson, Vincent Emanuele
Adversarial Attack Adversarial Example Reverse Engineering Attack Algorithm Reverse Noise

November 12, 2022

Generating Textual Adversaries with Minimal Perturbation
Xingyi Zhao, Lu Zhang, Depeng Xu, Shuhan Yuan
Adversarial Text Small Perturbation Attack Algorithm Word Level Adversarial

September 7, 2022

Defending Against Backdoor Attack on Graph Nerual Network by Explainability
Bingchen Jiang, Zhao Li
Backdoor Attack High Explainability Backdoor Detection Attack Algorithm Trojaned Model

August 21, 2022

Byzantines can also Learn from History: Fall of Centered Clipping in Federated Learning
Kerem Ozfatura, Emre Ozfatura, Alptekin Kupcu, Deniz Gunduz
Historical Text State of the Art Defense Byzantine Attack Malicious Client High Dimensional Clipping Byzantine Seal Attack Algorithm Robust Aggregator

June 20, 2022

Diversified Adversarial Attacks based on Conjugate Gradient Method
Keiichiro Yamamura, Haruki Sato, Nariaki Tateiwa, Nozomi Hata, Toru Mitsutake, Issa Oe, Hiroki Ishikura, Katsuki Fujisawa
Adversarial Attack Adversarial Example Conjugate Gradient Attack Algorithm

March 29, 2022

Recent improvements of ASR models in the face of adversarial attacks
Raphael Olivier, Bhiksha Raj
Adversarial Attack Native Robustness Recent Advance Human Face ASR Model Speech Recognition Model Attack Algorithm

January 10, 2022

IoTGAN: GAN Powered Camouflage Against Machine Learning Based IoT Device Identification
Tao Hou, Tao Wang, Zhuo Lu, Yao Liu, Yalin Sagduyu
Machine Learning Adversarial Perturbation Adversarial CAmouflage Attack Algorithm IoT Device Identification

Attack Algorithm

Papers

Leveraging Label Information for Stealthy Data Stealing in Vertical Federated Learning

PRP: Propagating Universal Perturbations to Attack Large Language Model Guard-Rails

Time-Distributed Backdoor Attacks on Federated Spiking Learning

Preference Poisoning Attacks on Reward Model Learning

Spoofing attack augmentation: can differently-trained attack models improve generalisation?

Adversarial Ink: Componentwise Backward Error Attacks on Deep Learning

Reliable and Efficient Evaluation of Adversarial Robustness for Deep Hashing-Based Retrieval

Reverse engineering adversarial attacks with fingerprints from adversarial examples

Generating Textual Adversaries with Minimal Perturbation

Defending Against Backdoor Attack on Graph Nerual Network by Explainability

Byzantines can also Learn from History: Fall of Centered Clipping in Federated Learning

Diversified Adversarial Attacks based on Conjugate Gradient Method

Recent improvements of ASR models in the face of adversarial attacks

IoTGAN: GAN Powered Camouflage Against Machine Learning Based IoT Device Identification