Iclr2020: Compression based bound for non-compressed network: unified generalization error analysis of large compressible deep neural network

Iclr2020: Compression based bound for non-compressed network: unified generalization error analysis of large compressible deep neural network - Download as a PDF or view online for free
1) The document presents a new compression-based bound for analyzing the generalization error of large deep neural networks, even when the networks are not explicitly compressed. 2) It shows that if a trained network's weights and covariance matrices exhibit low-rank properties, then the network has a small intrinsic dimensionality and can be efficiently compressed. 3) This allows deriving a tighter generalization bound than existing approaches, providing insight into why overparameterized networks generalize well despite having more parameters than training examples.

ICLR 2020

Iclr2020: Compression based bound for non-compressed network: unified generalization error analysis of large compressible deep neural network

Entropy, Free Full-Text

Taiji Suzuki, 准教授 at University of tokyo

Iclr2020: Compression based bound for non-compressed network: unified generalization error analysis of large compressible deep neural network

Entropy, Free Full-Text

PDF) MICIK: MIning Cross-Layer Inherent Similarity Knowledge for Deep Model Compression

PDF) Efficient Visual Recognition with Deep Neural Networks: A Survey on Recent Advances and New Directions

ICLR 2020

ICLR 2020