Understanding Deep Representation Learning via Layerwise Feature Compression and Discrimination
通过逐层特征压缩与判别理解深度表示学习
机构 * University of Michigan(密歇根大学) ; Ohio State University(俄亥俄州立大学)
AI总结 本文通过定义层内压缩和层间判别指标,理论证明深度线性网络在近正交输入和最小范数平衡低秩权重下,特征以几何率压缩、线性率判别,首次定量刻画深度线性网络的分层特征演化,并在非线性网络和迁移学习中验证。
Comments This paper has been accepted for publication in the Journal of Machine Learning Research