首页 时政热点 科技头条 智能AI 安全攻防 数码硬件 开发者生态 归档 标签 关于
智能AI morning

S-GAI:S形 MLP 的谱几何感知初始化——从数据集几何到网络权重

2026-06-30 1 阅读 约3分钟阅读 Yi-Shan Chu
分享:
字号:
arXiv:2606.28444v1 Announce Type: new Abstract: Classical universal approximation theorems establish the expressive power of sigmoidal multilayer perceptrons, but they do not prescribe how initial weights should encode the geometry of a data distribution.我们提出了 S-GAI,一种用于单隐藏层 sigmoidal MLP 的光谱几何感知初始化框架。 Starting from the constructive idea that sigmoid units can act as smooth half-space gates, we move from hand-specified planar geometry to class-wise spectral geometry estimated from image data.对于每个类别,SVD 提供平均值、主方向和光谱尺度。能量阈值选择保留方向,每个保留方向由两个 sigmoid 门表示。这些特定于类的门形成直接从训练集初始化的共享隐藏层。 We also formulate a SVD-based subspace classifier as a non-neural geometric reference, which tests whether the estimated spectral class geometry is already discriminative before being embedded into the MLP. Experiments on MNIST, Fashion-MNIST, and a more challenging CIFAR-10 test show that the S-GAI-initialized MLP starts from a substantially more informative hidden state than Xavier initialization and reaches comparable final accuracy under full training.当隐藏层被冻结时,仅训练输出层仍然比冻结随机门提供更强的性能,这提供了 S-GAI 有效地将按类谱几何嵌入到 MLP 中的证据。
这篇文章对您有帮助吗?

订阅66必读

每日精选科技资讯,直达你的邮箱