Adversarial Examples that Fool both Computer Vision and Time-Limited Humans

Abstract

Machine learning models are vulnerable to adversarial examples: small changes to images can cause computer vision models to make mistakes such as identifying a school bus as an ostrich. However, it is still an open question whether humans are prone to similar mistakes. Here, we address this question by leveraging recent techniques that transfer adversarial examples from computer vision models with known parameters and architecture to other models with unknown parameters and architecture, and by matching the initial processing of the human visual system. We find that adversarial examples that strongly transfer across computer vision models influence the classifications made by time-limited human observers.

Cited by in corpus (41)

• On Evaluating Adversarial Robustness
• A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications
• Motivating the Rules of the Game for Adversarial Example Research
• The Origins and Prevalence of Texture Bias in Convolutional Neural Networks
• Controversial stimuli: pitting neural networks against each other as models of human recognition
• A Survey on AI Assurance
• When Machine Learning Meets Privacy: A Survey and Outlook
• Generalizing semi-supervised generative adversarial networks to regression using feature contrasting
• Robust in Practice: Adversarial Attacks on Quantum Machine Learning
• Neural Networks with Recurrent Generative Feedback
• Can Adversarial Network Attack be Defended?
• Biologically Inspired Mechanisms for Adversarial Robustness
• Using learned optimizers to make models robust to input noise
• Dual Attention Suppression Attack: Generate Adversarial Camouflage in Physical World
• Two Souls in an Adversarial Image: Towards Universal Adversarial Example Detection using Multi-view Inconsistency
• Adversarial attacks hidden in plain sight
• Adversarial Examples on Object Recognition: A Comprehensive Survey
• If deep learning is the answer, then what is the question?
• Identifying Classes Susceptible to Adversarial Attacks
• The Intriguing Relation Between Counterfactual Explanations and Adversarial Examples
• Transforming Neural Network Visual Representations to Predict Human Judgments of Similarity
• Rearchitecting Classification Frameworks For Increased Robustness
• PhysGAN: Generating Physical-World-Resilient Adversarial Examples for Autonomous Driving
• Machine-learning Based Extraction of the Short-Range Part of the Interaction in Non-contact Atomic Force Microscopy
• Simple Transparent Adversarial Examples
• Optimizing Privacy-Preserving Outsourced Convolutional Neural Network Predictions
• AF2-Mutation: Adversarial Sequence Mutations against AlphaFold2 on Protein Tertiary Structure Prediction
• Representation of White- and Black-Box Adversarial Examples in Deep Neural Networks and Humans: A Functional Magnetic Resonance Imaging Study
• Seeing in the dark with recurrent convolutional neural networks
• Recurrent Attention Model with Log-Polar Mapping is Robust against Adversarial Attacks
• ImageNet-trained deep neural network exhibits illusion-like response to the Scintillating Grid
• Immuno-mimetic Deep Neural Networks (Immuno-Net)
• Can you hear me $\text{now}$? Sensitive comparisons of human and machine perception
• Energy Attack: On Transferring Adversarial Examples
• The Selectivity and Competition of the Mind's Eye in Visual Perception
• Attack Type Agnostic Perceptual Enhancement of Adversarial Images
• Computational principles of intelligence: learning and reasoning with neural networks
• Defense Through Diverse Directions
• PDPGD: Primal-Dual Proximal Gradient Descent Adversarial Attack
• Achilles Heels for AGI/ASI via Decision Theoretic Adversaries
• Evaluating adversarial robustness in simulated cerebellum