AI Paper Atlas

Research map / Continual learning / Online and replay-based

A Faster Path to Continual Learning

Wei Li, Hangjie Yuan, Zixiang Zhao, Borui Kang, Ziwei Liu, Tao Feng

CVPR 2026 · Continual learning › Online and replay-based · approach: Unified and unsupervised continual frameworks · 0 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 C-Flat with C-Flat TurboEASE · CIFAR100 B0_Inc10average accuracy: 91.91 -> 92.36 (C-Flat 92.05)Improved
replace C-Flat with C-Flat TurboEASE · CIFAR100 B0_Inc10throughput (% of vanilla): 26.5% -> 61.6%Improved
replace C-Flat with C-Flat TurboL2P · CIFAR100 B0_Inc10throughput (% of vanilla): 30.0% -> 59.4%Improved
add C-Flat Turbo instead of C-FlatMEMO · CIFAR100 from scratch ResNet-18last accuracy: 30.76 -> 32.24Improved
add C-Flat Turbo instead of C-FlatiCaRL · CIFAR100 from scratch ResNet-18last accuracy: 42.47 -> 42.84Improved
add SAM to the base optimizerL2P · CIFAR100 B0_Inc10average accuracy: 87.92 -> 88.18No clear effect
add LookSAM to the base optimizerEASE · CIFAR100 B0_Inc10average accuracy: 91.16 -> 90.89Hurt
add the reuse of cached flatness direction (ablation)EASE+C-Flat · CIFAR100 10-splitaverage accuracy: 92.15 -> 92.08; Img/s 44.25 -> 67.20Mixed
add linear step-size scheduler (ablation)EASE+C-Flat Turbo reuse · CIFAR100 10-splitthroughput (Img/s): 67.20 -> 74.63Improved
add adaptive trigger (ablation)EASE+C-Flat Turbo · CIFAR100 10-splitaverage accuracy: 92.03 -> 92.25; Img/s 74.63 -> 102.74Improved
increase trigger threshold m from 1 to 2 (ablation)C-Flat Turbo · CIFAR100 EASEaverage accuracy: 92.06 -> 91.86; Img/s 102.74 -> 154.52Hurt
use trigger m=0.1 instead of 1 (ablation)C-Flat Turbo · CIFAR100 EASEthroughput (Img/s): 102.74 -> 75.56Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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