Get Started
Home
Topics
Search
Library
Research questionHow can black-box adversarial attacks disrupt image-based continuous-control agents with fewer environment queries?In image-based continuous control, an attacker must infer how visual changes affect an agent while querying the environment. Large action spaces make this inference especially expensive and can weaken attacks.
AI
Alignment & Safety
Computer Vision
Machine Learning
Reinforcement Learning
Latest papersRecent research connected to this question, newest first.SEBA: Sample-Efficient Black-Box Attacks on Visual Reinforcement LearningApplies to black-box attacks on visual reinforcement-learning agents, including image-based continuous-control settings. The source reports evidence from MuJoCo and Atari benchmarks and considers visually imperceptible perturbations; conclusions are limited to the evaluated benchmark environments.research paper · Sep 2, 2026
Related questions
How can deployed vision models resist diverse backdoor triggers without model internals or auxiliary data?How can image-poisoning attacks remain effective across supervised and self-supervised learning without obvious visual artifacts?How can adversarial perturbations transfer to unseen semantic-segmentation models while accounting for dense spatial and class-wise structure?How can one vision model remain robust across changing and unseen adversarial perturbation budgets?