AI Paper Atlas

Research map / Continual learning / Online and replay-based

Optimizing Reusable Knowledge for Continual Learning via Metalearning

Julio Hurtado, Alain Raymond, Alvaro Soto

NeurIPS 2021 · Continual learning › Online and replay-based · approach: Theoretical analysis of forgetting · 3 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 metalearning KB update to baseline (ablation)sequential baseline · 20-Split CIFAR-100forgetting: -3% v/s -0.25%Improved
add mask-generating functions to Baseline + ML (full MARK) (ablation)Baseline + ML · 20-Split CIFAR-100forgetting (BWT): -3 -> -0.25Improved
add metalearning to Baseline + Mask (full MARK) (ablation)Baseline + Mask · 20-Split CIFAR-100max average accuracy: 74 -> 78Improved
add mask functions only to baseline (ablation)sequential baseline · 20-Split CIFAR-100accuracy and forgettingMixed
add metalearning only to baseline (ablation)sequential baseline · 20-Split CIFAR-100forgettingImproved
retrain each task after sequential learning (ablation)MARK-Task · 20-Split CIFAR-100task accuracy: +1.24%Improved
retrain vs maximum accuracy during training (ablation)MARK-Task · 20-Split CIFAR-100task accuracy: +0.97%Improved
use pretrained ImageNet ResNet-18 features instead of per-task featuresMARK · 20-Split CIFAR-100accuracy: 78.31 -> 86.29Improved
use random weights for feature embedding instead of trained (ablation)MARK · 20-Split MiniImageNetaccuracy: 69.43 -> 33.02Hurt
MARK-Task vs ACL accuracyMARK-Task · 20-Split MiniImageNetaccuracy: 62.07 -> 69.43Improved
MARK-Task vs ACL memoryMARK-Task · 20-Split CIFAR-100memory: 134% -> 100%Improved
remove KB mask retraining (No-Retraining variant) (ablation)MARK · 20-Split MiniImageNettask accuracyHurt

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