AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty

ICLR 2020 Dan HendrycksNorman MuEkin D. CubukBarret ZophJustin GilmerBalaji Lakshminarayanan

Modern deep neural networks can achieve high accuracy when the training distribution and test distribution are identically distributed, but this assumption is frequently violated in practice. When the train and test distributions are mismatched, accuracy can plummet... (read more)

PDF Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Domain Generalization ImageNet-C AugMix (ResNet-50) mean Corruption Error (mCE) 65.3 # 2
Domain Generalization ImageNet-R AugMix (ResNet-50) Top-1 Error Rate 58.9 # 4

Methods used in the Paper