Joint Adversarial

Joint adversarial techniques explore how to improve robustness and generalization in machine learning models by strategically incorporating adversarial examples into the training process. Current research focuses on developing novel adversarial attacks targeting specific model vulnerabilities, such as those affecting image generation or robotic control, and creating robust defenses using methods like adversarial training, incorporating diverse data augmentations (e.g., mixup strategies), and exploiting multi-modal information. This work is significant for enhancing the reliability and security of AI systems across various applications, from image synthesis and robotics to person re-identification, by mitigating the impact of malicious or unexpected inputs.

Papers

February 6, 2024

FoolSDEdit: Deceptively Steering Your Edits Towards Targeted Attribute-aware Distribution
Qi Zhou, Dongxia Wang, Tianlin Li, Zhihong Xu, Yang Liu, Kui Ren, Wenhai Wang, Qing Guo
Adversarial Attack Adversarial Noise Deceptive Diffusion Generative Attack Blur Attack Attribute Distribution Joint Adversarial

September 11, 2023

Revisiting Energy Based Models as Policies: Ranking Noise Contrastive Estimation and Interpolating Energy Models
Sumeet Singh, Stephen Tu, Vikas Sindhwani
Generative Modeling Behavior Cloning Energy Based Model Prior Policy Contrastive Estimation Energy Modeling Policy Representation Joint Adversarial

July 15, 2023

Cross-Model Cross-Stream Learning for Self-Supervised Human Action Recognition
Mengyuan Liu, Hong Liu, Tianyu Guo
Contrastive Learning Collaborative Learning Mutual Learning Self Supervised Action Joint Adversarial

May 20, 2022

Adversarial joint attacks on legged robots
Takuto Otomo, Hiroshi Kera, Kazuhiko Kawamoto
Adversarial Attack Adversarial Perturbation Legged Robot Bipedal Robot Multiple Gait Joint Adversarial

December 12, 2021

Interpolated Joint Space Adversarial Training for Robust and Generalizable Defenses
Chun Pong Lau, Jiang Liu, Hossein Souri, Wei-An Lin, Soheil Feizi, Rama Chellappa
Adversarial Attack Adversarial Training Joint Adversarial Manifold Attack

November 18, 2021

Person Re-identification Method Based on Color Attack and Joint Defence
Yunpeng Gong, Liqing Huang, Lifei Chen
Person Re Identification Image Fusion Joint Adversarial Adversarial Disturbance