paper-with-me

홈 › Papers

Low-Rank Curvature for Zeroth-Order Optimization in LLM Fine-Tuning

2025-11-11 · Hyunseok Seung, Jaewoo Lee, Hyunsuk Ko arxiv

We introduce LOREN, a curvature-aware zeroth-order (ZO) optimization method for fine-tuning large language models (LLMs). Existing ZO methods, which estimate gradients via finite differences using random perturbations, often suffer from high variance and suboptimal search directions. Our approach addresses these challenges by: (i) reformulating the problem of gradient preconditioning as that of adaptively estimating an anisotropic perturbation distribution for gradient estimation, (ii) capturing curvature through a low-rank block diagonal preconditioner using the framework of natural evolution strategies, and (iii) applying a REINFORCE leave-one-out (RLOO) gradient estimator to reduce variance. Experiments on standard LLM benchmarks show that our method outperforms state-of-the-art ZO methods by achieving higher accuracy and faster convergence, while cutting peak memory usage by up to 27.3% compared with MeZO-Adam.

📄 PDF Abstract BibTeX arXiv:2511.07971

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

CurvZO: Adaptive Curvature-Guided Sparse Zeroth-Order Optimization for Efficient LLM Fine-Tuning

2026-03-23 · Shuo Wang, Ziyu Chen, Ming Tang arxiv

Fine-tuning large language models (LLMs) with backpropagation achieves high performance but incurs substantial memory overhead, limiting scalability on resource-constrained hardware. Zeroth-order (ZO) optimization provid…

Second-Order Fine-Tuning without Pain for LLMs:A Hessian Informed Zeroth-Order Optimizer

2024-02-23 · Yanjun Zhao, Sizhe Dang, Haishan Ye, Guang Dai 외

Fine-tuning large language models (LLMs) with classic first-order optimizers entails prohibitive GPU memory due to the backpropagation process. Recent works have turned to zeroth-order optimizers for fine-tuning, which s…

GPU

Zeroth-Order Negative Curvature Finding: Escaping Saddle Points without Gradients

2022-10-04 · Hualin Zhang, Huan Xiong, Bin Gu

We consider escaping saddle points of nonconvex problems where only the function evaluations can be accessed. Although a variety of works have been proposed, the majority of them require either second or first-order info…

Reconciling Hessian-Informed Acceleration and Scalar-Only Communication for Efficient Federated Zeroth-Order Fine-Tuning

2025-06-03 · Zhe Li, Bicheng Ying, Zidong Liu, Chaosheng Dong 외

Recent dimension-free communication frameworks in Federated Learning (FL), such as DeComFL, significantly reduce per-round communication by transmitting only scalars via zeroth-order stochastic gradient descent (ZO-SGD).…

Federated Learning

AR1-ZO: Topology-Aware Rank-1 Zeroth-Order Queries for High-Rank LoRA Fine-Tuning

2026-05-19 · Ziye Chen, Hongbin Lin, Chenyu Zhang, Xiangda Yan 외 arxiv

Zeroth-order (ZO) optimization enables large-language-model fine-tuning without storing backpropagation activations, while LoRA supplies compact trainable adapters. Combining them creates a rank paradox: increasing LoRA …