A non-Euclidean dual gradient ascent for entropically regularized SDPs is shown to converge with dimension-independent rates, achieving Sinkhorn-like complexity for optimal transport and optimal-scaling results for permutation synchronization SDPs.
The multiplicative weights update method: a meta-algorithm and applications.Theory of computing, 8(1):121–164, 2012
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
math.OC 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
Non-Euclidean dual gradient ascent for entropically regularized linear and semidefinite programming
A non-Euclidean dual gradient ascent for entropically regularized SDPs is shown to converge with dimension-independent rates, achieving Sinkhorn-like complexity for optimal transport and optimal-scaling results for permutation synchronization SDPs.