Français(FR) Broken scale symmetries in undercomplete linear autoencoders

随机梯度下降破坏线性自编码器中的尺度对称性，偏好大的解码器权重

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-05 04:00

研究人员在欠完备线性自编码器中发现了一种现象，其中有限步长的随机梯度下降（SGD）会破坏尺度对称性。此过程偏好主成分分析（PCA）解流形上的大解码器权重，导致定向尺度漂移。虽然这种漂移在解析上是可处理的，但它最终会遇到稳定性边界，导致根据某些度量标准，解具有更尖锐的特征，尽管其他尖锐度度量可能会朝相反方向移动。 AI

影响提供了对SGD如何影响模型几何的理论理解，可能为未来的模型设计和训练提供信息。

排序理由详细介绍机器学习理论新发现的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

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

本文如何被排名

Signal score: 6 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 详细介绍机器学习理论新发现的学术论文。[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: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.LG TIER_1 Français(FR) · Farhad Pashakhanloo, Jacob A. Zavatone-Veth · 2026-10-05 04:00

欠完备线性自编码器中被破坏的尺度对称性

arXiv:2610.03640v1 Announce Type: new Abstract: Neural network loss landscapes have many symmetries, which are preserved by gradient flow but broken by finite-stepsize stochastic gradient descent (SGD). A canonical example of such a symmetry is scale in homogeneous networks: one …

本文如何被排名

报道来源 [1]

欠完备线性自编码器中被破坏的尺度对称性

相关实体

相关话题