BriefGPT.xyz
Jun, 2024
高维空间中学习稀疏特征的最优修剪
Pruning is Optimal for Learning Sparse Features in High-Dimensions
HTML
PDF
Nuri Mert Vural, Murat A. Erdogdu
TL;DR
通过在高维度中训练剪枝神经网络并与梯度下降算法结合,我们研究了剪枝网络对广泛类统计模型学习的影响,发现剪枝神经网络在样本复杂度上相比未剪枝网络有提升,并引入了相关统计查询下界来支持这一观点。
Abstract
While it is commonly observed in practice that
pruning networks
to a certain level of
sparsity
can improve the quality of the features, a theoretical explanation of this phenomenon remains elusive. In this work,
→