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.
Future Internet, Free Full-Text
Iclr2020: Compression based bound for non-compressed network: unified generalization error analysis of large compressible deep neural network
Continuum Modeling and Control of Large Nonuniform Networks
ICLR 2020 Statistics - Paper Copilot
Intellectual property protection of DNN models
Publications - OATML
Publications - OATML
Adversarial Neural Pruning with Latent Vulnerability Suppression
Minimax optimal alternating minimization \ for kernel nonparametric tensor learning
ICLR 2020
Single-model uncertainty quantification in neural network potentials does not consistently outperform model ensembles
Peter Richtarik
Adversarial Neural Pruning with Latent Vulnerability Suppression
Machine learning project