AI Paper Atlas

Research map / Continual learning / Online and replay-based

Elastic Weight Consolidation Done Right for Continual Learning

Xuan Liu, Xiaobin Chang

CVPR 2026 · Continual learning › Online and replay-based · approach: Unified and unsupervised continual frameworks · 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 Logits Reversal when computing the Fisher importance (EWC to EWC-DR)EWC · CIFAR-100 EFCIL big-start T=5A_last: 14.61 -> 50.23Improved
add Logits Reversal when computing the Fisher importanceEWC · ImageNet-Subset EFCIL big-start T=5A_last: 11.44 -> 66.18Improved
add Logits Reversal when computing the Fisher importanceEWC · Tiny-ImageNet EFCIL big-start T=5A_last: 9.74 -> 38.24Improved
add Logits Reversal when computing the Fisher importanceEWC · CIFAR-100 EFCIL equally split T=20A_last: 5.48 -> 18.00Improved
add Logits Reversal when computing the Fisher importanceEWC · MCIT NLVR2 after final taskforgetting transfer (%): 90.66 -> 27.48Improved
add Logits Reversal when computing the Fisher importanceEWC · MCIT after final task VCRincremental accuracy A_t (%): 42.99 -> 52.59Improved
add Logits Reversal when computing the Fisher importanceEWC · MCIT new-task accuracy (plasticity)diagonal accuracy: VCR 59.30 -> 59.38No clear effect
accumulate importance online with decay (online EWC) instead of vanilla EWCEWC · CIFAR-100 EFCIL big-start T=5A_last: 14.61 -> 29.70Improved
use online EWC instead of vanilla EWC in equally-split settingEWC · ImageNet-Subset EFCIL equally split T=10A_last: 10.60 -> 12.78Mixed
use Synaptic Intelligence instead of vanilla EWCEWC · ImageNet-Subset EFCIL big-start T=5A_last: 11.44 -> 9.19Hurt
use MAS importance instead of vanilla EWCEWC · CIFAR-100 EFCIL big-start T=5A_last: 14.61 -> 35.37Improved

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