Derives accessible O(κΦ ln(κΦ ||w*||/ε)) iteration bound for rPDHG on unique-optima LPs, with computable Φ, two-stage performance, and equivalence to stability and sharpness.
arXiv preprint arXiv:2407.19689
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 5roles
background 1polarities
background 1representative citing papers
DSPDHG extends PDHG and SPDHG with doubly stochastic block updates and proves O(1/K) ergodic convergence for the expected restricted primal-dual gap plus linear convergence for a restarted variant under quadratic growth.
cuRegOT is a new GPU solver for entropic OT that delivers speedups over prior GPU methods via amortized analysis, asynchronous iterates, and fused kernels, backed by convergence guarantees.
An inexact projected-gradient framework for GWOT with a verifiable feasibility-residual condition that proves subsequential and full-sequence convergence to stationary points under a mild tolerance decay.
LAMP reduces optimal transport storage to linear space O(n+m) while preserving last-iterate convergence rates of primal-dual mirror prox and scaling to n=m=2^18.
citing papers explorer
-
Accessible Complexity Bounds for Restarted PDHG on Linear Programs with a Unique Optimizer
Derives accessible O(κΦ ln(κΦ ||w*||/ε)) iteration bound for rPDHG on unique-optima LPs, with computable Φ, two-stage performance, and equivalence to stability and sharpness.
-
On the convergence of doubly stochastic Primal-Dual Hybrid Gradient Method
DSPDHG extends PDHG and SPDHG with doubly stochastic block updates and proves O(1/K) ergodic convergence for the expected restricted primal-dual gap plus linear convergence for a restarted variant under quadratic growth.
-
cuRegOT: A GPU-Accelerated Solver for Entropic-Regularized Optimal Transport
cuRegOT is a new GPU solver for entropic OT that delivers speedups over prior GPU methods via amortized analysis, asynchronous iterates, and fused kernels, backed by convergence guarantees.
-
A Provably Convergent and Practical Algorithm for Gromov--Wasserstein Optimal Transport
An inexact projected-gradient framework for GWOT with a verifiable feasibility-residual condition that proves subsequential and full-sequence convergence to stationary points under a mild tolerance decay.
-
Log-Averaged Mirror Prox for Fast, Large-Scale Optimal Transport in Linear Space
LAMP reduces optimal transport storage to linear space O(n+m) while preserving last-iterate convergence rates of primal-dual mirror prox and scaling to n=m=2^18.