AI Paper Atlas

Research map / Continual learning / Class-incremental

Flexi-FSCIL: Adaptive Knowledge Retention for Breaking the Stability-Plasticity Dilemma in Few-Shot Class-Incremental Learning

Wufei Xie, Yalin Wang, Chenliang Liu, Zhaohui Jiang, Xue Yang

ICCV 2025 · Continual learning › Class-incremental · approach: Few-shot class-incremental prototypes · 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 AGRF (gated frozen+trainable fusion) to the base model (ablation)Flexi-FSCIL · CUB200accuracy delta vs no strategy: +34.08Improved
add ADHD distillation on top of AGRF (ablation)Flexi-FSCIL · CUB200last session accuracy (%): 52.77 -> 55.39Improved
add unlabeled data without POE filtering (ablation)Flexi-FSCIL · CUB200last session accuracy (%): 55.39 -> 58.32Improved
add POE selection of unlabeled samples (ablation)Flexi-FSCIL · CUB200last session accuracy (%): 58.32 -> 66.24Improved
add both FA and SP criteria in POE (ablation)Flexi-FSCIL · CUB200last session accuracy (%): 59.51 -> 66.24Improved
use AvgPool instead of spatial, channel or flatten similarity in ADHD (ablation)Flexi-FSCIL · CUB200PD (performance drop): 14.09 -> 12.97Improved
increase number of unlabeled samples beyond 160 (to 200) (ablation)Flexi-FSCIL · CUB200last session accuracy (%): 66.24 -> 64.02Hurt
use POE with 40 unlabeled samples vs without POE (ablation)Flexi-FSCIL · CUB200last session accuracy (%): 56.93 -> 63.19Improved
embed Flexi-FSCIL into FACLFACL · CUB200PD (performance drop): 18.04 -> 16.84Improved
embed Flexi-FSCIL into FACTFACT · CUB200PD (performance drop): 21.55 -> 20.32Improved
Flexi-FSCIL vs best baseline on miniImageNetFlexi-FSCIL · miniImageNetdelta F: +58.29 vs +58.08No clear effect

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