A one-parameter scaling law models excess loss from data repetition as an additive overfitting penalty, recommending model capacity increases over excessive repetition and showing that strong weight decay reduces the penalty coefficient by ~70%.
The Thirty-ninth Annual Conference on Neural Information Processing Systems , year=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Prescriptive Scaling Laws for Data Constrained Training
A one-parameter scaling law models excess loss from data repetition as an additive overfitting penalty, recommending model capacity increases over excessive repetition and showing that strong weight decay reduces the penalty coefficient by ~70%.