Depth expansion in normalized residual networks yields provable test-risk improvement through representational, optimization, and generalization gains under first-order descent and norm-control conditions.
Thus, if ∆ERM ≥2(ϵ M +ϵ K), then Ltest(fnew)≤ L test(f ∗ old)
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
-
A Qualitative Test-Risk Mechanism for Scaling Behavior in Normalized Residual Networks
Depth expansion in normalized residual networks yields provable test-risk improvement through representational, optimization, and generalization gains under first-order descent and norm-control conditions.