Stable rank normalization for improved generalization in neural networks and GANs

Stable rank normalization for improved generalization in neural networks and GANs

Exciting new work on generalization bounds for neural networks (NN) given by Bartlett et al. (2017); Neyshabur et al. (2018) closely depend on two parameter- dependant quantities: the Lipschitz constant upper bound and the stable rank (a softer version of rank). Even though these bounds typically ha...

書目詳細資料
Main Authors:	Sanyal, A, Torr, P, Dokania, P
格式:	Conference item
語言:	English
出版:	International Conference on Learning Representations 2020

相似書籍

On using focal loss for neural network calibration
由: Mukhoti, J, et al.
出版: (2020)

Calibrating deep neural networks using focal loss
由: Mukhoti, J, et al.
出版: (2020)

On Generalization Bounds for Neural Networks with Low Rank Layers
由: Pinto, Andrea, et al.
出版: (2024)

Proximal mean-field for neural network quantization
由: Ajanthan, T, et al.
出版: (2020)

Are vision transformers always more robust than convolutional neural networks?
由: Pinto, F, et al.
出版: (2021)

Mix-MaxEnt: improving accuracy and uncertainty estimates of deterministic neural networks
由: Pinto, F, et al.
出版: (2021)

Continual learning in low-rank orthogonal subspaces
由: Chaudhry, A, et al.
出版: (2020)

How benign is benign overfitting?
由: Sanyal, A, et al.
出版: (2021)

Mirror Descent view for Neural Network quantization
由: Ajanthan, T, et al.
出版: (2021)

Progressive skeletonization: trimming more fat from a network at initialization
由: de Jorge, P, et al.
出版: (2020)

Lyapunov-stable neural-network control
由: Dai, Hongkai, et al.
出版: (2022)

Improving generalization of neural network using length as discriminant
由: Siraj, Fadzilah, et al.
出版: (1999)

Normal Modified Stable Processes.
由: Shephard, N, et al.
出版: (2001)

Normal modified stable processes
由: Shephard, N, et al.
出版: (2001)

Neural SDEs as infinite-dimensional GANs
由: Kidger, P, et al.
出版: (2021)

Rough Neural Networks Architecture For Improving Generalization In Pattern Recognition
由: Ali Adlan, Hanan Hassan
出版: (2004)

Improving generalization of neural networks using multilayer perceptron discriminants
由: Siraj, Fadzilah, et al.
出版: (2002)

Speeding up convolutional neural networks with low rank expansions
由: Jaderberg, M, et al.
出版: (2014)

Speeding up convolutional neural networks with low rank expansions
由: Jaderberg, M, et al.
出版: (2014)

On the robustness of quality measures for GANs
由: Alfarra, M, et al.
出版: (2022)

Realistic face generation using deep neural networks (StyleGAN)
由: Toh, Wilson Chin Shen
出版: (2021)

On the existence of generalized rank weights
由: Oggier, Frederique, et al.
出版: (2013)

Generalized Low-Rank Approximations
由: Srebro, Nathan, et al.
出版: (2004)

What makes and breaks safety fine-tuning? a mechanistic study
由: Jain, S, et al.
出版: (2024)

Stable Rank-Adaptive Dynamically Orthogonal Runge–Kutta Schemes
由: Charous, Aaron, et al.
出版: (2024)

Varsities improve in rank
由: Menon, Sandhya
出版: (2021)

Varsities Improve On Rankings
由: The Star, Choi Tuck Wo
出版: (2008)

Stable homotopy and generalized homology/
由: 447926 Adams, John Frank
出版: (1974)

SGD Noise and Implicit Low-Rank Bias in Deep Neural Networks
由: Galanti, Tomer, et al.
出版: (2022)

Certifying ensembles: a general certification theory with s-lipschitzness
由: Petrov, A, et al.
出版: (2023)

Interpolating convolutional neural networks using batch normalization
由: Data, G, et al.
出版: (2018)

Multi-agent diverse generative adversarial networks
由: Ghosh, A, et al.
出版: (2018)

Stable and efficient computation of generalized polar decompositions
由: Benner, P, et al.
出版: (2022)

On Neural Network Pruning’s Effect on Generalization
由: Jin, Tian
出版: (2023)

On the Expressiveness and Generalization of Hypergraph Neural Networks
由: Luo, Zhezheng
出版: (2023)

Learning and generalization : with applications to neural networks /
由: 378518 Vidyasagar, M. Mathukumalli), 1947-
出版: (2003)

Development of low normal force contact with stable performance
由: Bhuiyan, Moinul
出版: (2010)

GeoNet++: Iterative geometric neural network with edge-aware refinement for joint depth and surface normal estimation
由: Qi, X, et al.
出版: (2020)

Interpreting neural network judgments via minimal, stable, and symbolic corrections
由: Solar Lezama, Armando, et al.
出版: (2021)

Improving generalization of neural networks using MLP discriminant based on multiple classifiers failures
由: Siraj, Fadzilah, et al.
出版: (2010)