Establishes explicit finite-sample bias and variance bounds for regularized OT costs that improve prior entropic results, deliver the first quantitative bounds for L^p regularization, and yield an n^{-2/(d+4)} rate for quadratic regularization with quadratic cost.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.ST 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Finite-sample bounds for regularized optimal transport
Establishes explicit finite-sample bias and variance bounds for regularized OT costs that improve prior entropic results, deliver the first quantitative bounds for L^p regularization, and yield an n^{-2/(d+4)} rate for quadratic regularization with quadratic cost.