AI Paper Atlas

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

Effective Meta-Regularization by Kernelized Proximal Regularization

Weisen Jiang, James Kwok, Yu Zhang

NeurIPS 2021 · Learning theory and algorithms › Kernels and regression theory · approach: High-dimensional asymptotic generalization · 1 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
add learnable proximal regularizer f_theta (MetaProx vs MetaOptNet-RR, f_theta=0)MetaOptNet-RR · Sine 2-shot noisyMSE: 0.79 -> 0.43Improved
add learnable f_thetaMetaOptNet-RR · Sine 2-shot noise-freeMSE: 0.18 -> 0.11Improved
add learnable f_thetaMetaOptNet-RR · Sine 5-shot noisyMSE: 0.19 -> 0.13Improved
add learnable f_thetaMetaOptNet-RR · Sine 5-shot noise-freeMSE: 0.01 -> 0.01No clear effect
add learnable f_thetaMetaOptNet-RR · Sale 1-shotMSE: 0.088 -> 0.061Improved
add learnable f_thetaMetaOptNet-RR · QMUL in-rangeMSE: 0.021 -> 0.012Improved
add learnable f_thetaMetaOptNet-RR · QMUL out-of-rangeMSE: 0.039 -> 0.020Improved
add learnable regularizer f_theta to MetaOptNet-SVM(cos), Conv4MetaOptNet-SVM(cos) · mini-ImageNet 5-way 1-shot Conv4accuracy: 50.1 -> 52.4Improved
add learnable regularizer f_theta to MetaOptNet-SVM(cos), ResNet-12MetaOptNet-SVM(cos) · mini-ImageNet 5-way 1-shot ResNet-12accuracy: 62.75 -> 63.82Improved
add learnable regularizer f_theta to MetaOptNet-SVM(cos), ResNet-12MetaOptNet-SVM(cos) · mini-ImageNet 5-way 5-shot ResNet-12accuracy: 78.68 -> 79.12Improved
use kernel-based methods instead of non-kernel meta-regularization (Meta-MinibatchProx)MetaProx · QMUL in-rangeMSE: 0.171 -> 0.012Improved
replace linear model CommonMean with kernel/neural methodsMetaProx · Sine 2-shot noise-freeMSE: 4.58 -> 0.11Improved

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