A hyperparameter-free gradient autoscaling method that zero-centers gradients and multiplies them by a global factor based on gradient standard deviation improves CIFAR-100 accuracy slightly on ResNets, but with weak statistical and theoretical support.
Gradnorm: Gra- dient normalization for adaptive loss balancing in deep multitask networks
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
REJECT 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Insights from Gradient Dynamics: Gradient Autoscaled Normalization
A hyperparameter-free gradient autoscaling method that zero-centers gradients and multiplies them by a global factor based on gradient standard deviation improves CIFAR-100 accuracy slightly on ResNets, but with weak statistical and theoretical support.