Adversary Knowledge

Adversary knowledge, in the context of machine learning, explores how an attacker's understanding of a target model influences the effectiveness of adversarial attacks—malicious inputs designed to cause misclassification or other failures. Current research focuses on characterizing the impact of different levels of adversary knowledge (e.g., white-box vs. black-box access) across various model architectures, including deep neural networks, graph neural networks, and reinforcement learning agents, and developing frameworks to systematically analyze these attacks. Understanding adversary knowledge is crucial for improving the robustness and security of machine learning systems in diverse applications, from network security to autonomous systems and data privacy.

Papers

May 2, 2024

Obtaining physical layer data of latest generation networks for investigating adversary attacks
M. V. Ushakova, Yu. A. Ushakov, L. V. Legashev
Machine Learning Machine Learning Model Machine Learning Application Massive MIMO Physical Layer Future Network Adversary Knowledge

February 22, 2024

SoK: Analyzing Adversarial Examples: A Framework to Study Adversary Knowledge
Lucas Fenaux, Florian Kerschbaum
Adversarial Example New Framework Threat Model Transferable Attack Adversary Knowledge

August 21, 2023

Stabilizing Unsupervised Environment Design with a Learned Adversary
Ishita Mediratta, Minqi Jiang, Jack Parker-Holder, Michael Dennis, Eugene Vinitsky, Tim Rocktäschel
Reinforcement Learning Procedural Content Generation Maze Navigation Unsupervised Environment Design Robust Agent Adversary Knowledge

July 27, 2023

FLARE: Fingerprinting Deep Reinforcement Learning Agents using Universal Adversarial Masks
Buse G. A. Tekgul, N. Asokan
Deep Reinforcement Learning Adversarial Example Retrieval Augmented Generation Adversarial Mask Adversary Knowledge DNN Fingerprint Removal Attack

October 13, 2022

Observed Adversaries in Deep Reinforcement Learning
Eugene Lim, Harold Soh
Adversarial Attack Deep Reinforcement Learning Adversarial Perspective Monotone Adversary Deep Policy Adversary Knowledge

September 2, 2022

Group Property Inference Attacks Against Graph Neural Networks
Xiuling Wang, Wendy Hui Wang
Graph Neural Network Privacy Attack Property Inference Attack Graph Level Adversary Knowledge

January 13, 2022

Reconstructing Training Data with Informed Adversaries
Borja Balle, Giovanni Cherubin, Jamie Hayes
Training Data Reconstruction Attack High Fidelity Reconstruction Monotone Adversary Adversary Knowledge

November 8, 2021

Characterizing the adversarial vulnerability of speech self-supervised learning
Haibin Wu, Bo Zheng, Xu Li, Xixin Wu, Hung-yi Lee, Helen Meng
Adversarial Attack Self Supervised Learning Speech Analysis Speech Representation Adversarial Vulnerability Speech Supervised Learning Model Zero Knowledge Adversary Knowledge