Based Attack

Based attacks exploit vulnerabilities in machine learning models to manipulate their outputs, aiming to either misclassify inputs (e.g., through adversarial patches or crafted prompts) or infer sensitive information about the training data (membership inference attacks). Current research focuses on developing and evaluating these attacks across various model types, including deep neural networks for image classification and large language models for text processing, employing techniques like gradient-based methods, quantile regression, and self-prompt calibration. Understanding and mitigating these attacks is crucial for ensuring the reliability and security of machine learning systems in diverse applications, from autonomous vehicles to privacy-preserving AI.

Papers

June 25, 2024

CoSafe: Evaluating Large Language Model Safety in Multi-Turn Dialogue Coreference
Erxin Yu, Jing Li, Ming Liao, Siqi Wang, Zuchen Gao, Fei Mi, Lanqing Hong
Large Language Model Secure Approach Prompt Attack Based Attack Dialogue Coreference

April 5, 2024

Evaluating Adversarial Robustness: A Comparison Of FGSM, Carlini-Wagner Attacks, And The Role of Distillation as Defense Mechanism
Trilokesh Ranjan Sarkar, Nilanjan Das, Pralay Sankar Maitra, Bijoy Some, Ritwik Saha, Orijita Adhikary, Bishal Bose, Jaydip Sen
Deep Neural Network Adversarial Attack Adversarial Robustness Mutual Distillation DEfense Mechanism Fast Gradient Sign Method Based Attack Defensive Distillation

March 8, 2024

EVD4UAV: An Altitude-Sensitive Benchmark to Evade Vehicle Detection in UAV
Huiming Sun, Jiacheng Guo, Zibo Meng, Tianyun Zhang, Jianwu Fang, Yuewei Lin, Hongkai Yu
Unmanned Aerial Vehicle UAV Image Based Attack

November 10, 2023

Practical Membership Inference Attacks against Fine-tuned Large Language Models via Self-prompt Calibration
Wenjie Fu, Huandong Wang, Chen Gao, Guanghua Liu, Yong Li, Tao Jiang
Language Model Membership Inference Attack Prompt Tuning Fine Tuned Large Language Model Based Attack

August 25, 2023

Active learning for fast and slow modeling attacks on Arbiter PUFs
Vincent Dumoulin, Wenjing Rao, Natasha Devroye
Active Learning Attack Model Unclonable Function Cryptographic Primitive Based Attack

July 7, 2023

Scalable Membership Inference Attacks via Quantile Regression
Martin Bertran, Shuai Tang, Michael Kearns, Jamie Morgenstern, Aaron Roth, Zhiwei Steven Wu
Black Box Membership Inference Attack Membership Inference Based Attack

May 29, 2023

Membership Inference Attacks against Language Models via Neighbourhood Comparison
Justus Mattern, Fatemehsadat Mireshghallah, Zhijing Jin, Bernhard Schölkopf, Mrinmaya Sachan, Taylor Berg-Kirkpatrick
Language Model Adversarial Attack Training Data Membership Inference Attack Neighborhood Level Based Attack

November 3, 2022

Data-free Defense of Black Box Models Against Adversarial Attacks
Gaurav Kumar Nayak, Inder Khatri, Ruchit Rawal, Anirban Chakraborty
Adversarial Attack Black Box Model Based Attack

August 18, 2022

Shadows Aren't So Dangerous After All: A Fast and Robust Defense Against Shadow-Based Adversarial Attacks
Andrew Wang, Wyatt Mayor, Ryan Smith, Gopal Nookula, Gregory Ditzler
Adversarial Attack Quantum Shadow Robust Defense Traffic Sign Recognition Based Attack

March 22, 2022

Semi-Targeted Model Poisoning Attack on Federated Learning via Backward Error Analysis
Yuwei Sun, Hideya Ochiai, Jun Sakuma
Poisoning Attack Model Poisoning Attack Attack Performance Based Attack

Based Attack

Papers

CoSafe: Evaluating Large Language Model Safety in Multi-Turn Dialogue Coreference

Evaluating Adversarial Robustness: A Comparison Of FGSM, Carlini-Wagner Attacks, And The Role of Distillation as Defense Mechanism

EVD4UAV: An Altitude-Sensitive Benchmark to Evade Vehicle Detection in UAV

Practical Membership Inference Attacks against Fine-tuned Large Language Models via Self-prompt Calibration

Active learning for fast and slow modeling attacks on Arbiter PUFs

Scalable Membership Inference Attacks via Quantile Regression

Membership Inference Attacks against Language Models via Neighbourhood Comparison

Data-free Defense of Black Box Models Against Adversarial Attacks

Shadows Aren't So Dangerous After All: A Fast and Robust Defense Against Shadow-Based Adversarial Attacks

Semi-Targeted Model Poisoning Attack on Federated Learning via Backward Error Analysis