adversarial example
PulseAugur coverage of adversarial example — every cluster mentioning adversarial example across labs, papers, and developer communities, ranked by signal.
-
Federated Learning faces new adversarial attack and defense research · 2 sources tracked
Two recent arXiv papers explore the challenges of adversarial attacks and defenses within federated learning (FL) frameworks. The first paper investigates the feasibility of adversarial training for Vision Transformers …
-
Federated Learning Faces New Adversarial Attacks and Defenses
This paper explores the vulnerabilities of federated learning (FL) systems to various adversarial attacks, including poisoning, Byzantine, and adversarial example attacks. Researchers analyzed the transferability of adv…
-
New Perspective-Invariant Attack Enhances Adversarial Example Transferability
Researchers have developed a new method called Perspective-Invariant Attack (PIA) to enhance the transferability of adversarial examples in deep neural networks. Unlike previous methods that used limited local transform…
-
AI-generated image detectors vulnerable to adversarial attacks
Researchers have discovered that reconstruction-based detectors, designed to identify AI-generated images without training, are vulnerable to adversarial attacks. These attacks manipulate images to artificially increase…
-
Security threats and defenses for embodied AI with world models surveyed
This paper surveys the security landscape of embodied AI systems that utilize world models. It details how these predictive cores, while enabling advanced planning, also introduce new vulnerabilities. The research trace…
-
New physical attack method targets optical flow estimation networks with infrared lights
Researchers have developed a novel method to physically attack Optical Flow Estimation Networks (OFENs) in real-time using infrared lights. This approach generates numerous adversarial examples in advance and displays t…
-
New framework uses causal reasoning to detect adversarial AI examples
Researchers have introduced CausAdv, a novel framework designed to detect adversarial examples in deep learning models, particularly Convolutional Neural Networks (CNNs) used in computer vision. This approach leverages …
-
New defense Random Logit Scaling protects AI models from adversarial attacks
Researchers have introduced Random Logit Scaling (RLS), a new defense mechanism designed to protect deep neural networks against black-box score-based adversarial example attacks. RLS functions as a post-processing step…
-
New framework unifies detection of AI content, hallucinations, and watermarks
Researchers have developed a novel unified framework for detecting AI-generated content and artifacts, including LLM text, hallucinations, watermarks, and adversarial examples. The method utilizes Mahalanobis distance s…
-
New research targets AI robustness with novel distillation and testing methods · 8 sources tracked
Researchers are exploring new methods to enhance the adversarial robustness of neural networks. One approach, AD-CERT, combines adversarial distillation with Interval Bound Propagation to achieve state-of-the-art certif…