DRIFT preprocesses images by projecting them onto fixed sine mode shapes, which lets small networks train with tens of features while showing smoother convergence than PCA or pixel inputs.
Adversarially robust generalization requires more data
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
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
DRIFT: Data Reduction via Informative Feature Transformation- Generalization Begins Before Deep Learning starts
DRIFT preprocesses images by projecting them onto fixed sine mode shapes, which lets small networks train with tens of features while showing smoother convergence than PCA or pixel inputs.