S³LDBO is a snapshot single-loop algorithm for decentralized bilevel optimization that reduces computational cost via intermittent derivative skipping and provides ergodic and high-probability nonergodic iteration complexity bounds in deterministic settings.
On the convergence analysis of the decentralized projected gradient descent,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.OC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
S$^3$LDBO: A Snapshot Single-Loop Algorithm for Decentralized Bilevel Optimization
S³LDBO is a snapshot single-loop algorithm for decentralized bilevel optimization that reduces computational cost via intermittent derivative skipping and provides ergodic and high-probability nonergodic iteration complexity bounds in deterministic settings.