CORTEXA
← Browse
arxivcs.LGmath.OC2026-07-24

On the Convergence of Stochastic Low-Rank Adaptation

Ru Wang, Chengchang Liu, John C. S. Lui

Low-rank adaptation (LoRA) optimizes $J(B,A)=\mathcal L(W_\mathrm{base}+sBA)$ over two adapters $B \in \mathbb{R}^{m \times r}$ and $A \in \mathbb{R}^{r \times n}$ that form a low-rank update to a frozen pretrained weight matrix $W_\mathrm{base} \in \mathbb{R}^{m \times n}$. The prior analysis shows LoRA-GD takes $\exp\{\mathcal{O}(ε^{-2})\}$ oracle calls to find an $ε$-stationary point such that $\|\nabla J(B,A)\|\leq ε$ in the deterministic setting. We sharpen the analysis and show that $\mathcal{O}(ε^{-4})$ full-gradient evaluations suffice for the same first-order criterion. We further study stochastic LoRA under unbiased gradient estimates and finite variance. We propose LoRA-NSGDM, which finds an $ε$-stationary point with $\mathcal{O}(ε^{-8})$ stochastic oracle complexity. Under the additional mean-square smoothness condition, we use variance reduction strategy and propose LoRA-STORM, which improves the stochastic oracle complexity to $\mathcal{O}(ε^{-6})$.

View free PDFSource page

Related papers

arxivmath.OCcs.LG2026-06-30

Direction-Magnitude Decomposition for Low-Rank Matrix Optimization: Faster Convergence and Saddle-to-saddle Dynamics

Yudong Wei, Liang Zhang, Bingcong Li, Niao He

Low-rank matrix optimization is often carried out via the Burer-Monteiro (BM) formulation, but choosing the factorization rank $r$ is delicate and can substantially slow optimization. We propose a unified framework, termed direction-magnitude decomposition (DMD), that decomposes…

View free PDFSource page
arxivcs.LGmath.OC2026-07-01

ZO-Act: Efficient Zeroth-Order Fine-Tuning via One-Shot Activation-Informed Low-Rank Subspaces

Xun Dong, Yibo Xu, Naigang Wang, Xin Li, Penghang Yin, Zi Yang

Zeroth-order (ZO) optimization enables fine-tuning large language models when backpropagation is unavailable or memory-prohibitive, but existing methods often perturb full model weights or randomly constructed low-dimensional subspaces, yielding high-variance estimates and limite…

View free PDFSource page
arxivcs.LGmath.NAmath.OC2026-07-10

Graph-Regularized Low-Rank Matrix Completion by Variable Projection

Benoît Loucheur, P. -A. Absil, Michel Journée

We address the low-rank matrix completion problem by incorporating graph regularization into the existing Riemannian Trust-Region Matrix Completion (RTRMC) framework. The latter uses the geometry of the low-rank constraint to remodel the problem as an unconstrained optimization p…

View free PDFSource page
arxivcs.LGmath.OC2026-07-16

Regularity-Aware Stochastic MGDA with Adaptive Conflict-Avoidant Update Direction Control

Chentong Huang, Lisha Chen

Multi-objective learning (MOL) aims to optimize multiple objectives simultaneously. The multi-gradient descent algorithm (MGDA) is a workhorse that iteratively updates along a common descent or conflict-avoidant (CA) direction across objectives. In stochastic settings, however, t…

View free PDFSource page