Physic Based Adversarial

Physics-based adversarial methods focus on creating and mitigating adversarial examples that leverage physical principles, improving the robustness and reliability of machine learning models. Current research explores generating adversarial examples by simulating physical processes like camera malfunctions or material stress, and using physics-informed training to enhance model accuracy and consistency with known physical laws. These techniques are particularly relevant for safety-critical applications such as autonomous driving and aircraft systems, where model trustworthiness is paramount, and for improving the efficiency of scientific simulations by generating synthetic data. The overarching goal is to develop more reliable and physically consistent machine learning models across various domains.

Papers

July 24, 2024

Physical Adversarial Attack on Monocular Depth Estimation via Shape-Varying Patches
Chenxing Zhao, Yang Li, Shihao Wu, Wenyi Tan, Shuangju Zhou, Quan Pan
Adversarial Attack Monocular Depth Estimation Adversarial Patch Attack Class Relevant Patch Depth Estimation Error Physic Based Adversarial

May 23, 2024

Generating camera failures as a class of physics-based adversarial examples
Manav Prabhakar, Jwalandhar Girnar, Arpan Kusari
Adversarial Example Adversarial Sample True Class Physical Simulation Stress Distribution Physic Based Adversarial

March 28, 2023

Generating artificial digital image correlation data using physics-guided adversarial networks
David Melching, Erik Schultheis, Eric Breitbarth
Generative Adversarial Network Displacement Field Digital Image Correlation Physic Based Adversarial

September 7, 2022

Physics-Guided Adversarial Machine Learning for Aircraft Systems Simulation
Houssem Ben Braiek, Thomas Reid, Foutse Khomh
Adversarial Learning Physic Based Adversarial

July 14, 2022

PIAT: Physics Informed Adversarial Training for Solving Partial Differential Equations
Simin Shekarpaz, Mohammad Azizmalayeri, Mohammad Hossein Rohban
Adversarial Attack Adversarial Training Partial Differential Equation Theoretical Physic Adversarial Discriminator Physic Based Adversarial

Physic Based Adversarial

Papers

Physical Adversarial Attack on Monocular Depth Estimation via Shape-Varying Patches

Generating camera failures as a class of physics-based adversarial examples

Generating artificial digital image correlation data using physics-guided adversarial networks

Physics-Guided Adversarial Machine Learning for Aircraft Systems Simulation

PIAT: Physics Informed Adversarial Training for Solving Partial Differential Equations