Revealing Modular Gradient Noise Imbalance in LLMs: Calibrating Adam via Signal-to-Noise Ratio

• URL: http://arxiv.org/abs/2605.05794v1
• Date: Thu, 07 May 2026 07:32:27 GMT
• Title: Revealing Modular Gradient Noise Imbalance in LLMs: Calibrating Adam via Signal-to-Noise Ratio
• Abstract summary: We introduce textbfModule-wise Learning Rate Scaling via SNR (MoLS).<n>MoLS estimates module-level SNRs to scale Adam updates, allowing automated module-wise learning rate allocation without manual tuning.<n> Empirical results through multiple LLM training benchmarks demonstrate that MoLS improves convergence speed and generalization, achieving performance comparable to carefully tuned module-specific learning rates.
• Score: 45.81412107070137
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The impressive performance of large language models (LLMs) arises from their massive scale and heterogeneous module composition. However, this structural heterogeneity introduces additional optimization challenges. While adaptive optimizers such as Adam(W) provide per-parameter adaptivity, they do not explicitly account for module-level gradient heterogeneity, resulting in slower convergence, suboptimal performance, or training instability. Existing approaches typically rely on manually tuned module-specific learning rates or specific optimization strategies, which are computationally costly and difficult to generalize across tasks or models. To establish a more principled approach, we first analyze the noise-damping behavior of Adam in high-noise modules and introduce \textbf{Module-wise Learning Rate Scaling via SNR (MoLS)}. MoLS estimates module-level SNRs to scale Adam updates, allowing automated module-wise learning rate allocation without manual tuning. Empirical results through multiple LLM training benchmarks demonstrate that MoLS improves convergence speed and generalization, achieving performance comparable to carefully tuned module-specific learning rates, while remaining compatible with memory-efficient training algorithms.

Related papers

• Adaptive Linear Path Model-Based Diffusion [52.84663832658799]
  We introduce Linear Path Model-Based Diffusion (LP-MBD), which replaces the variance-preserving schedule with a flow-matching linear probability path.<n>We also propose Adaptive LP-MBD (ALP-MBD), which leverages reinforcement learning to adjust diffusion steps and noise levels according to task complexity and environmental conditions.
  arXiv  Detail & Related papers  (2026-02-02T21:33:03Z)
• Taming LLMs by Scaling Learning Rates with Gradient Grouping [49.91587150497186]
  Training large language models (LLMs) poses challenges due to their massive scale and heterogeneous architectures.<n>This work introduces Scaling with Gradient Grouping (SGG), an gradient wrapper that improves adaptive learning rate estimation by dynamic grouping and group-specific scaling.
  arXiv  Detail & Related papers  (2025-06-01T15:30:37Z)
• Leveraging Importance Sampling to Detach Alignment Modules from Large Language Models [48.15777554876988]
  Traditional alignment methods often require retraining large pretrained models.<n>We propose a novel textitResidual Alignment Model (textitRAM) that formalizes the alignment process as a type of importance sampling.<n>We develop a resampling algorithm with iterative token-level decoding to address the common first-token latency issue in comparable methods.
  arXiv  Detail & Related papers  (2025-05-26T08:53:02Z)
• Zeroth-Order Fine-Tuning of LLMs in Random Subspaces [63.10833446782114]
  As language models grow in size, memory demands for backpropagation increase.<n>Zeroth-order (ZO) optimization methods offer a memory-efficient alternative.<n>In this paper, we propose Subspace Zero-order optimization to address the challenges posed by posed by high dimensionality perturbations.
  arXiv  Detail & Related papers  (2024-10-11T17:01:43Z)
• Duo-LLM: A Framework for Studying Adaptive Computation in Large Language Models [16.16372459671255]
  Large Language Models (LLMs) typically generate outputs token by token using a fixed compute budget. We propose a novel framework that integrates smaller auxiliary modules within each Feed-Forward Network layer of the LLM. We show that trained routers operate differently from oracles and often yield suboptimal solutions.
  arXiv  Detail & Related papers  (2024-10-01T16:10:21Z)
• Spectrum: Targeted Training on Signal to Noise Ratio [39.74810517186403]
  We present Spectrum, a method that accelerates LLM training by selectively targeting layer modules based on their signal-to-noise ratio (SNR) Our approach, which utilizes an algorithm to compute module SNRs prior to training, has shown to effectively match the performance of full fine-tuning while reducing GPU memory usage.
  arXiv  Detail & Related papers  (2024-06-07T21:20:57Z)
• Train Faster, Perform Better: Modular Adaptive Training in Over-Parameterized Models [31.960749305728488]
  We introduce a novel concept dubbed modular neural tangent kernel (mNTK) We show that the quality of a module's learning is tightly associated with its mNTK's principal eigenvalue $lambda_max$. We propose a novel training strategy termed Modular Adaptive Training (MAT) to update those modules with their $lambda_max$ exceeding a dynamic threshold.
  arXiv  Detail & Related papers  (2024-05-13T07:46:48Z)
• Meta-Causal Feature Learning for Out-of-Distribution Generalization [71.38239243414091]
  This paper presents a balanced meta-causal learner (BMCL), which includes a balanced task generation module (BTG) and a meta-causal feature learning module (MCFL) BMCL effectively identifies the class-invariant visual regions for classification and may serve as a general framework to improve the performance of the state-of-the-art methods.
  arXiv  Detail & Related papers  (2022-08-22T09:07:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.