PulseAugur
中
实时 08:17:37
English(EN) From Geometry to Generalization: Why Row Normalization Can Beat Adam and Muon

行归一化在高维分类中优于 Adam 和 Muon

一篇新的研究论文探讨了机器学习中优化器的几何特性,提出在हरूको维多类别分类任务中,行归一化可以优于 Adam 和 Muon。研究表明,与 Adam 的逐坐标几何和 Muon 的谱几何相比,行归一化的逐类欧几里得几何能更好地保持决策边界方向,从而避免失真。这一优势在包括各向同性高斯云数据和幂律谱在内的各种数据模型下得到了证明,并且通过合成和语言模型实验支持了这些发现。 AI

影响 提出了一种新的优化技术,可能带来更准确、更高效的机器学习模型。

排序理由 一篇发表在 arXiv 上的研究论文,详细介绍了一种新颖的优化器方法。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

行归一化在高维分类中优于 Adam 和 Muon

本文如何被排名

Signal score
18 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
一篇发表在 arXiv 上的研究论文,详细介绍了一种新颖的优化器方法。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
paper, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

  1. arXiv cs.AI TIER_1 English(EN) · Jihwan Kim, Dogyoon Song, Chulhee Yun ·

    从几何到泛化:为何行归一化能胜过 Adam 和 Muon

    arXiv:2610.11309v1 Announce Type: cross Abstract: Different optimizers can fit the same training data while selecting classifiers with substantially different geometries, but whether this difference provably affects population performance remains unclear. We show that row-wise no…