SAM solves a min-max problem to locate flat low-loss regions, improving generalization on CIFAR, ImageNet and label-noise tasks.
We can ensure thatj∈ N using inequality equation 7 and by setting c = ρ2(1 + exp(4n/k))
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2020 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Sharpness-Aware Minimization for Efficiently Improving Generalization
SAM solves a min-max problem to locate flat low-loss regions, improving generalization on CIFAR, ImageNet and label-noise tasks.