Standard online learning of functions with bounded derivative integral on R permits infinite adversarial loss, but distance-restricted or weighted scenarios yield finite sharp bounds for fast-decaying weights in 1D while remaining ill-posed in d>=2.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
Online learning of smooth functions on $\mathbb{R}$
Standard online learning of functions with bounded derivative integral on R permits infinite adversarial loss, but distance-restricted or weighted scenarios yield finite sharp bounds for fast-decaying weights in 1D while remaining ill-posed in d>=2.