AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

Adaptive Curriculum Learning

Yajing Kong, Liu Liu, Jun Wang, Dacheng Tao

ICCV 2021 · Label-efficient and robust learning › Knowledge distillation · approach: Teacher properties and robustness · 53 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 difficulty-score adaptation (with KL loss) (ablation)Adaptive CL · CIFAR-100validation accuracy: 45.49 -> 46.08Improved
add KL divergence loss (with adaptation) (ablation)Adaptive CL · CIFAR-100validation accuracy: 45.88 -> 46.08Improved
add difficulty-score adaptation (with KL loss) (ablation)Adaptive CL · HE devicesvalidation accuracy: 72.25 -> 74.06Improved
add KL divergence loss (with adaptation) (ablation)Adaptive CL · HE devicesvalidation accuracy: 73.64 -> 74.06Improved
replace fixed curriculum with Adaptive CLFCL · CIFAR-100validation accuracy: 45.25 -> 46.08Improved
replace random sampling with Adaptive CLVanilla · CIFAR-100validation accuracy: 42.16 -> 46.08Improved
use self-paced learning instead of vanilla samplingVanilla · CIFAR-100validation accuracy: 42.16 -> 39.70Hurt
use focal loss instead of vanilla samplingVanilla · CIFAR-100validation accuracy: 42.16 -> 42.34No clear effect
use VGG-16-based initial difficulty score: Adaptive CL instead of FCLFCL · HE devicesvalidation accuracy: 73.32 -> 74.17Improved
use VGG-16-based initial difficulty score: Adaptive CL instead of FCLFCL · MS mammalsvalidation accuracy: 75.34 -> 75.84Improved
use VGG-16-based initial difficulty score: Adaptive CL instead of FCLFCL · Flowersvalidation accuracy: 68.90 -> 69.45Improved
use negative alpha (adaptation) instead of alpha=0 (ablation)Adaptive CL w/o KL · HE devicesvalidation accuracy: about 2%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