Norm-matched zeroth-order adaptation preserves the isotropic retention floor while contracting only the anisotropic component, producing a quadratic forgetting gap that favors ZO precisely when the first-order direction has above-average retention curvature.
Learning multiple layers of features from tiny images
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
-
Why Zeroth-Order Adaptation May Forget Less: A Randomized Shaping Theory
Norm-matched zeroth-order adaptation preserves the isotropic retention floor while contracting only the anisotropic component, producing a quadratic forgetting gap that favors ZO precisely when the first-order direction has above-average retention curvature.