HORST uses non-commutative operator composition and a hyperbolic mirror map to combine stability from adaptive optimizers with L1 sparsity bias, outperforming AdamW across sparsity levels on vision and language tasks.
Title resolution pending
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 4verdicts
UNVERDICTED 4roles
background 1polarities
background 1representative citing papers
PowerStep delivers coordinate-wise adaptive optimization by nonlinearly transforming a momentum buffer under an lp-norm steepest-descent geometry, matching Adam convergence with half the memory and supporting aggressive quantization.
Three-average primal-dual methods achieve accelerated rates for computable accuracy certificates in convex optimization.
The paper motivates stochastic optimization problems from statistical perspectives and describes offline and online approaches to solve expectation minimization problems.
citing papers explorer
-
HORST: Composing Optimizer Geometries for Sparse Transformer Training
HORST uses non-commutative operator composition and a hyperbolic mirror map to combine stability from adaptive optimizers with L1 sparsity bias, outperforming AdamW across sparsity levels on vision and language tasks.
-
PowerStep: Memory-Efficient Adaptive Optimization via $\ell_p$-Norm Steepest Descent
PowerStep delivers coordinate-wise adaptive optimization by nonlinearly transforming a momentum buffer under an lp-norm steepest-descent geometry, matching Adam convergence with half the memory and supporting aggressive quantization.
-
Accuracy Certificates for Convex Optimization at Accelerated Rates via Primal-Dual Averaging
Three-average primal-dual methods achieve accelerated rates for computable accuracy certificates in convex optimization.
-
Stochastic Optimization and Data Science
The paper motivates stochastic optimization problems from statistical perspectives and describes offline and online approaches to solve expectation minimization problems.