AI Paper Atlas

Research map / Continual learning / Online and replay-based

A continual learning survey: Defying forgetting in classification tasks

Matthias Delange, Rahaf Aljundi, Marc Masana, Sarah Parisot, Xu Jia, Aleš Leonardis, Greg Slabaugh, Tinne Tuytelaars

TPAMI 2021 · Continual learning › Online and replay-based · approach: Theoretical analysis of forgetting · 2,075 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 iCaRL instead of basic rehearsal with full memoryreplay with 4.5k buffer · Tiny ImageNet (BASE)average accuracy: 37.31 -> 47.27Improved
use GEM instead of basic rehearsal with partial memoryreplay with 4.5k buffer · Tiny ImageNet (BASE)average accuracy: 36.09 -> 45.13Improved
double replay buffer from 4.5k to 9k (ablation)iCaRL · Tiny ImageNet (BASE)average accuracy: 47.27 -> 48.76Improved
double replay buffer from 4.5k to 9k (ablation)GEM · Tiny ImageNet (BASE)average accuracy: 45.13 -> 41.75Hurt
add feature-autoencoder (EBLL) to LwFLwF · Tiny ImageNet (BASE)average accuracy: 41.91 -> 45.34Improved
use MAS instead of EWCregularization-based CL · Tiny ImageNet (BASE)average accuracy: 42.43 -> 46.90Improved
use DEEP model instead of BASE (ablation)PackNet · Tiny ImageNet (random order)average accuracy: 49.13 -> 35.54Hurt
use WIDE model instead of BASE (ablation)EWC · Tiny ImageNet (random order)average accuracy: 42.43 -> 31.10Hurt
use SMALL model instead of BASE (ablation)EWC · Tiny ImageNet (random order)average accuracy: 42.43 -> 45.13Improved
add dropout (p=0.5) (ablation)finetuning · Tiny ImageNet (BASE)average accuracy: 21.30 -> 29.23Improved
add dropout (p=0.5) (ablation)PackNet · Tiny ImageNet (BASE)average accuracy: 47.67 -> 54.28Improved
add dropout (p=0.5) (ablation)SI · Tiny ImageNet (BASE)average accuracy: 33.93 -> 43.15Improved

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