Connecting the Dots: Detecting Adversarial Perturbations Using Context Inconsistency
arXiv:2007.09763
Abstract
There has been a recent surge in research on adversarial perturbations that defeat Deep Neural Networks (DNNs) in machine vision; most of these perturbation-based attacks target object classifiers. Inspired by the observation that humans are able to recognize objects that appear out of place in a scene or along with other unlikely objects, we augment the DNN with a system that learns context consistency rules during training and checks for the violations of the same during testing. Our approach builds a set of auto-encoders, one for each object class, appropriately trained so as to output a discrepancy between the input and output if an added adversarial perturbation violates context consistency rules. Experiments on PASCAL VOC and MS COCO show that our method effectively detects various adversarial attacks and achieves high ROC-AUC (over 0.95 in most cases); this corresponds to over 20% improvement over a state-of-the-art context-agnostic method.
The paper is accepted by ECCV 2020
References in corpus (10)
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- Improving neural networks by preventing co-adaptation of feature detectors
- Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks
- Synthesizing Robust Adversarial Examples
- Detecting Adversarial Samples from Artifacts
- Early Methods for Detecting Adversarial Images
- Characterizing Adversarial Examples Based on Spatial Consistency Information for Semantic Segmentation
- MMM: Multi-stage Multi-task Learning for Multi-choice Reading Comprehension
- Identifying and Resisting Adversarial Videos Using Temporal Consistency
- A4 : Evading Learning-based Adblockers