A kernel-regression iteration over forward and reverse simulated paths computes Schrödinger bridge potentials with provable, minimax-optimal convergence rates.
An automorphism of product measures.Ann
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Forward Reverse Kernel Regression for the Schr\"{o}dinger bridge problem
A kernel-regression iteration over forward and reverse simulated paths computes Schrödinger bridge potentials with provable, minimax-optimal convergence rates.