Skip to content
Hot eventLive

Muon优化器两带谱重加权框架BulkBoost研究

1 reports1 sources3 hr ago updated

Get the story

AI overview

2026年10月7日,arXiv Statistics Machine Learning 发表一手研究,提出 BulkBoost 双带谱重加权框架,探讨 Muon 优化器是否需要细粒度谱形塑造。该框架将低于 Marchenko–Pastur 噪声边的 bulk 与高于边的 spike 分开,仅靠一个共享的 bulk-to-spike 增益调整权重,并给出固定秩与噪声校准两种变体。目前报道仅涉及方法框架,未见实验结果或后续验证。

Generated from reports · updated 3 hr ago

Timeline

Follow the coverage from different angles.

Oct 7, 2026
  1. arXiv · Statistics Machine Learning
    Muon 是否需要细粒度谱形塑造?

    研究提出 BulkBoost 双带谱重加权框架,用固定秩与噪声校准两种变体,把低于 Marchenko–Pastur 噪声边的 bulk 与高于边的 spike 分开,仅靠一个共享的 bulk-to-spike 增益调整权重。

Heat trend

Current heat 9·Comparable peak 10(Oct 7)·Comparable change over 24 hours –

02.557.510Oct7Oct7Oct7Oct7

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.