AI Paper Atlas

Research map / Learning theory and algorithms / Kernels and regression theory

Normalized Flat Minima: Exploring Scale Invariant Definition of Flat Minima for Neural Networks Using PAC-Bayesian Analysis

Yusuke Tsuzuku, Issei Sato, Masashi Sugiyama

ICML 2020 · Learning theory and algorithms › Kernels and regression theory · approach: Empirical generalization measures · 13 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use normalized sharpness instead of trace of the HessianWide ResNet · CIFAR-10, varying random-label ratioscorrelation coefficient with generalization gap: 70.8 -> 91.1Improved
Use normalized sharpness to distinguish models trained on random labelsMLP and LeNet · MNISTassociation between sharpness and generalization gapImproved
Use normalized sharpness to distinguish models trained on random labelsLeNet and Wide ResNet · CIFAR-10association between sharpness and generalization gapImproved
Remove scale normalization from the flatness measureFlatness measures on tested networks · MNIST and CIFAR-10ability to distinguish random-label modelsHurt
Use an alternative direct optimization definition of normalized sharpness instead of the second-order approximationNormalized sharpness calculation · Not evaluated experimentallyoptimization convexityHurt
Use an alternative direct optimization definition of normalized sharpness instead of the second-order approximationNormalized sharpness calculation · Not evaluated experimentallysensitivity to the choice of lambdaHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy