English
Related papers

Related papers: Zeroth-Order Optimization at the Edge of Stability

200 papers

Zeroth-order optimization (ZOO) is an important framework for stochastic optimization when gradients are unavailable or expensive to compute. A potential limitation of existing ZOO methods is the bias inherent in most gradient estimators…

Machine Learning · Computer Science 2025-10-24 Shaocong Ma , Heng Huang

We investigate the effectiveness of adaptive zeroth-order (ZO) optimization for memory-constrained fine-tuning of large language models (LLMs). Contrary to prior claims, we show that adaptive ZO methods such as ZO-Adam offer no convergence…

Machine Learning · Computer Science 2026-05-06 Hassan Dbouk , Nidham Gazagnadou , Matthias Reisser , Christos Louizos

$\ell_0$ constrained optimization is prevalent in machine learning, particularly for high-dimensional problems, because it is a fundamental approach to achieve sparse learning. Hard-thresholding gradient descent is a dominant technique to…

Machine Learning · Computer Science 2024-03-19 William de Vazelhes , Hualin Zhang , Huimin Wu , Xiao-Tong Yuan , Bin Gu

Large language models (LLMs) excel across various tasks, but standard first-order (FO) fine-tuning demands considerable memory, significantly limiting real-world deployment. Recently, zeroth-order (ZO) optimization stood out as a promising…

Machine Learning · Computer Science 2025-11-04 Qitao Tan , Jun Liu , Zheng Zhan , Caiwei Ding , Yanzhi Wang , Xiaolong Ma , Jaewoo Lee , Jin Lu , Geng Yuan

Classical optimization theory establishes that zeroth-order (ZO) algorithms suffer from a dimension-dependent slowdown, with convergence rates typically scaling with the model dimension compared to first-order methods. However, in contrast…

Machine Learning · Computer Science 2026-05-06 Zhe Li , Bicheng Ying , Zidong Liu , Haibo Yang

Fine-tuning large language models (LLMs) often faces GPU memory bottlenecks: the backward pass of first-order optimizers like Adam increases memory usage to more than 10 times the inference level (e.g., 633 GB for OPT-30B). Zeroth-order…

Machine Learning · Computer Science 2025-07-01 Sizhe Dang , Yangyang Guo , Yanjun Zhao , Haishan Ye , Xiaodong Zheng , Guang Dai , Ivor Tsang

In the evolving landscape of natural language processing (NLP), fine-tuning pre-trained Large Language Models (LLMs) with first-order (FO) optimizers like SGD and Adam has become standard. Yet, as LLMs grow {in size}, the substantial memory…

Zeroth-order (ZO) optimization is an emerging deep neural network (DNN) training paradigm that offers computational simplicity and memory savings. However, this seemingly promising approach faces a significant and long-ignored challenge. ZO…

Machine Learning · Computer Science 2025-07-28 Qitao Tan , Sung-En Chang , Rui Xia , Huidong Ji , Chence Yang , Ci Zhang , Jun Liu , Zheng Zhan , Zhenman Fang , Zhou Zou , Yanzhi Wang , Jin Lu , Geng Yuan

Safe derivative-free optimization under unknown constraints is a fundamental challenge in modern learning and control. Existing zeroth-order (ZO) methods typically still assume access to a first-order oracle of the constraint functions or…

Optimization and Control · Mathematics 2026-01-29 Runyu Zhang , Gioele Zardini , Asuman Ozdaglar , Jeff Shamma , Na Li

Zeroth-order (ZO) optimization with ordinal feedback has emerged as a fundamental problem in modern machine learning systems, particularly in human-in-the-loop settings such as reinforcement learning from human feedback, preference…

Optimization and Control · Mathematics 2025-12-23 Haishan Ye

Continual learning requires new-task adaptation without damaging previously acquired capabilities. Recent forward-pass and zeroth-order (ZO) results show that low-query adaptation may retain better than first-order (FO) descent, but the…

Machine Learning · Computer Science 2026-05-12 Yao Shu , Jian Mu , Zhongxiang Dai

Fine-tuning is powerful for adapting large language models to downstream tasks, but it often results in huge memory usages. A promising approach to mitigate this is using Zeroth-Order (ZO) optimization, which estimates gradients to replace…

Machine Learning · Computer Science 2024-10-15 Fei Wang , Li Shen , Liang Ding , Chao Xue , Ye Liu , Changxing Ding

We empirically demonstrate that full-batch gradient descent on neural network training objectives typically operates in a regime we call the Edge of Stability. In this regime, the maximum eigenvalue of the training loss Hessian hovers just…

Machine Learning · Computer Science 2022-11-24 Jeremy M. Cohen , Simran Kaur , Yuanzhi Li , J. Zico Kolter , Ameet Talwalkar

Zeroth-order (ZO) optimization is being recognized as a simple yet powerful alternative to standard backpropagation (BP)-based training. Notably, ZO optimization allows for training with only forward passes and (almost) the same memory as…

Machine Learning · Computer Science 2025-01-09 Keisuke Sugiura , Hiroki Matsutani

Optimizing large-scale nonconvex problems, common in deep learning, demands balancing rapid convergence with computational efficiency. First-order (FO) optimizers, which serve as today's baselines, provide fast convergence and good…

Machine Learning · Computer Science 2025-09-30 Jiahe Chen , Ziye Ma

Fine-tuning Large Language Models (LLMs) has proven effective for a variety of downstream tasks. However, as LLMs grow in size, the memory demands for backpropagation become increasingly prohibitive. Zeroth-order (ZO) optimization methods…

Machine Learning · Computer Science 2025-07-25 Ziming Yu , Pan Zhou , Sike Wang , Jia Li , Mi Tian , Hua Huang

Rank-based zeroth-order (ZO) optimization -- which relies only on the ordering of function evaluations -- offers strong robustness to noise and monotone transformations, and underlies many successful algorithms such as CMA-ES, natural…

Machine Learning · Computer Science 2025-12-19 Haishan Ye

Zeroth-order methods have become important tools for solving problems where we have access only to function evaluations. However, the zeroth-order methods only using gradient approximations are $n$ times slower than classical first-order…

Optimization and Control · Mathematics 2022-02-10 Erik Berglund , Sarit Khirirat , Xiaoyu Wang

Fine-tuning large language models (LLMs) has achieved remarkable success across various NLP tasks, but the substantial memory overhead during backpropagation remains a critical bottleneck, especially as model scales grow. Zeroth-order (ZO)…

Computation and Language · Computer Science 2026-01-09 Feihu Jin , Shipeng Cen , Ying Tan

We consider the problem of minimizing a high-dimensional objective function, which may include a regularization term, using (possibly noisy) evaluations of the function. Such optimization is also called derivative-free, zeroth-order, or…

Optimization and Control · Mathematics 2023-03-20 HanQin Cai , Daniel Mckenzie , Wotao Yin , Zhenliang Zhang