Publication

A jamming transition from under- to over-parametrization affects generalization in deep learning

Related publications (52)

About
Privacy
Disclaimer

Graph Chatbot

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.

A jamming transition from under- to over-parametrization affects generalization in deep learning

Graph Chatbot

Chat with Graph Search

SparseFool: a few pixels make a big difference

The committee machine: Computational to statistical gaps in learning a two-layers neural network

Empirical study of the topology and geometry of deep networks

Eigendecomposition-free Training of Deep Networks with Zero Eigenvalue-based Losses

Real-time DCT Learning-based Reconstruction of Neural Signals

Modeling Facial Geometry using Compositional VAEs

Modeling Facial Geometry using Compositional VAEs

Robustness of Classifiers to Universal Perturbations: A Geometric Perspective

Visual Analysis of Maya Glyphs via Crowdsourcing and Deep Learning

Adaptive data augmentation for image classification

Visual Analysis of Maya Glyphs via Crowdsourcing and Deep Learning

Robustness of Classifiers to Universal Perturbations: A Geometric Perspective

Real-time DCT Learning-based Reconstruction of Neural Signals

Empirical study of the topology and geometry of deep networks

SparseFool: a few pixels make a big difference

Adaptive data augmentation for image classification

The committee machine: Computational to statistical gaps in learning a two-layers neural network

Modeling Facial Geometry using Compositional VAEs

Modeling Facial Geometry using Compositional VAEs

Eigendecomposition-free Training of Deep Networks with Zero Eigenvalue-based Losses