LGMLSep 23, 2025

Functional Scaling Laws in Kernel Regression: Loss Dynamics and Learning Rate Schedules

arXiv:2509.19189v38 citationsh-index: 1
Originality Incremental advance
AI Analysis

This provides a theoretical framework for optimizing training efficiency in large-scale AI, though it is incremental by extending scaling laws to dynamics.

The paper tackles the problem of understanding loss dynamics and learning rate schedules in large language model training by establishing a Functional Scaling Law (FSL) for kernel regression, showing that higher-capacity models are more efficient and warmup-stable-decay schedules outperform pure decay, with experiments validating predictions on models up to 1B parameters.

Scaling laws have emerged as a unifying lens for understanding and guiding the training of large language models (LLMs). However, existing studies predominantly focus on the final-step loss, leaving open whether the entire loss dynamics obey similar laws and, crucially, how the learning rate schedule (LRS) shapes them. We address these gaps in a controlled theoretical setting by analyzing stochastic gradient descent (SGD) on a power-law kernel regression model. The key insight is a novel intrinsic-time viewpoint, which captures the training progress more faithfully than iteration count. We then establish a Functional Scaling Law (FSL) that captures the full loss trajectory under arbitrary LRSs, with the schedule's influence entering through a simple convolutional functional. We further instantiate the theory for three representative LRSs -- constant, exponential decay, and warmup-stable-decay (WSD) -- and derive explicit scaling relations in both data- and compute-limited regimes. These comparisons explain key empirical phenomena: (i) higher-capacity models are more data- and compute-efficient; (ii) learning-rate decay improves training efficiency; and (iii) WSD-type schedules outperform pure decay. Finally, experiments on LLMs ranging from 0.1B to 1B parameters demonstrate the practical relevance of FSL as a surrogate model for fitting and predicting loss trajectories in large-scale pre-training.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes