Adversarial Examples Improve Image Recognition

21 Nov 2019Cihang XieMingxing TanBoqing GongJiang WangAlan YuilleQuoc V. Le

Adversarial examples are commonly viewed as a threat to ConvNets. Here we present an opposite perspective: adversarial examples can be used to improve image recognition models if harnessed in the right manner... (read more)

PDF Abstract

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Image Classification ImageNet AdvProp (EfficientNet-B8) Top 1 Accuracy 85.5% # 5
Image Classification ImageNet AdvProp (EfficientNet-B8) Top 5 Accuracy 97.3% # 7
Image Classification ImageNet AdvProp (EfficientNet-B8) Number of params 88M # 1
Image Classification ImageNet AdvProp (EfficientNet-B7) Top 1 Accuracy 85.2% # 7
Image Classification ImageNet AdvProp (EfficientNet-B7) Top 5 Accuracy 97.2% # 8
Image Classification ImageNet AdvProp (EfficientNet-B7) Number of params 66M # 1