AI Paper Atlas

Research map / Optimization / Optimizers

On Modulating the Gradient for Meta-Learning

Christian Simon, Piotr Koniusz, Richard Nock, Mehrtash Harandi

ECCV 2020 · Optimization › Optimizers · approach: Gradient-based meta-learning theory · 49 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
replace CAVIA-style methods with ModGrad modulation (1-step Conv-4)CAVIA (512) · mini-ImageNet 5-way 1-shotaccuracy: 51.82 -> 53.20Improved
replace CAVIA-style methods with ModGrad modulation (1-step Conv-4)CAVIA (512) · mini-ImageNet 5-way 5-shotaccuracy: 65.85 -> 69.17Improved
replace Meta-Curvature with ModGrad on WRN-28-10Meta-Curvature · mini-ImageNet 5-way 1-shotaccuracy: 64.40 -> 65.72Improved
replace Meta-Curvature with ModGrad on WRN-28-10Meta-Curvature · mini-ImageNet 5-way 5-shotaccuracy: 80.21 -> 81.17Improved
replace LEAP with ModGrad on multi-shot OmniglotLEAP · Omniglot 20-way multi-shotaccuracy: 83.3 -> 85.1Improved
use ModGrad instead of CAVIA on image completion (10 random pixels)CAVIA · CelebA image completion (10 random pixels)accuracy: 0.037 -> 0.034Improved
run more inner-loop steps (5 instead of 1) in ModGrad (ablation)ModGrad · mini-ImageNet 5-way 5-shotaccuracy: 69.17 -> 68.24No clear effect
run more inner-loop steps (5 instead of 1) in ModGrad (ablation)ModGrad · mini-ImageNet 5-way 1-shotaccuracy: 53.20 -> 52.76No clear effect
increase column dimension u from 5 to 10 (ablation)ModGrad · mini-ImageNet 5-way 5-shotaccuracy: 69.17 -> 67.62Hurt
decrease column dimension u from 5 to 1 (ablation)ModGrad · mini-ImageNet 5-way 1-shotaccuracy: 53.20 -> 53.13No clear effect
use ModGrad with 1 step instead of MAML with 30 steps on ResNet-34 20-shotMAML · mini-ImageNet 5-way 20-shotaccuracy: 2%Improved
add ModGrad under gradient noiseMeta-SGD, CAVIA, T-Net, MAML · mini-ImageNet (noisy gradients)accuracy: 10%Improved

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