AI Paper Atlas

Research map / Continual learning / Class-incremental

Few-shot Class-Incremental Learning via Generative Co-Memory Regularization

Kexin Bao, Yong Li, Dan Zeng, Shiming Ge

IJCV 2026 · Continual learning › Class-incremental · approach: Few-shot class-incremental prototypes · 1 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 co-memory regularization to incremental learning with ResNet12 (vs C-FSCIL)ResNet12 FSCIL · CIFAR100 FSCILavg accuracy over sessions: 61.64 -> 67.65Improved
add co-memory regularization versus NC-FSCIL with ResNet12ResNet12 FSCIL · MiniImageNet FSCILavg accuracy over sessions: 67.82 -> 68.04Improved
add weight memory and representation memory together (beta 0.8) vs weight memory only (ablation)GCMR ViT-B · MiniImageNet FSCILavg accuracy over sessions: 66.84 -> 70.26Improved
use beta 1.0 (representation memory only) vs beta 0.0 (ablation)GCMR ViT-B · MiniImageNet FSCILavg accuracy over sessions: 66.84 -> 70.55Improved
combine memories with beta 0.7 vs weight memory only (ablation)GCMR ViT-B · CIFAR100 FSCILavg accuracy over sessions: 68.48 -> 69.77Improved
replace ResNet18 with ViT-B encoder at beta 0.3 (ablation)GCMR · MiniImageNet FSCILavg accuracy over sessions: 68.15 -> 68.15No clear effect
replace ResNet18 with ViT-B encoder at beta 0.3 (ablation)GCMR · CIFAR100 FSCILavg accuracy over sessions: 68.26 -> 69.12Improved
use ResNet34 instead of ResNet18 encoder (ablation)GCMR beta 0.3 · MiniImageNet FSCILavg accuracy over sessions: 68.15 -> 68.32Improved
add generative domain adaptation finetuning (ablation)GCMR base session · MiniImageNet/CIFAR100/CUB200base session accuracy: +0.47 to +1.49Improved
add pretraining of the ViT encoder (ablation)GCMR base session · MiniImageNetbase session accuracy: over 40% than that without pretrainingImproved
GCMR vs NC-FSCIL base session on CUB200GCMR · CUB200 FSCILbase session accuracy: 80.45 -> 80.64Improved
large pretrained models (PriViLege) vs GCMRPriViLege · CIFAR100 FSCILlast session accuracy: 58.84 -> 86.06Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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