AI Paper Atlas

Research map / Continual learning / Class-incremental

FLAR: A Unified Prototype Framework for Few-Sample Lifelong Active Recognition

Lei Fan, Peixi Xiong, Wei Wei, Ying Wu

ICCV 2021 · Continual learning › Class-incremental · approach: Class-balanced knowledge retention · 8 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
use learned prototype-based policy instead of random view selectionFLAR active recognition agent · SUN360 (26 categories)accuracy at t=5 (average): 62.3 -> 71.1Improved
use prototype reward and prototype representation instead of Look-Ahead rewardFLAR active recognition agent · SUN360 (26 categories)accuracy at t=2 (current): 59.8 -> 61.5Improved
use learned policy instead of largest-step movementFLAR active recognition agent · SUN360 (26 categories)accuracy at t=5 (current): 52.4 -> 69.9Improved
increase new-category training samples from k=5 to k=10 (ablation)FLAR · SUN360 class-incrementalaccuracyNo clear effect
increase new-category training samples from k=3 to k=5 (ablation)FLAR · SUN360 class-incrementalaccuracyImproved
vary stored exemplars m=1,3,5 (ablation)FLAR · SUN360 class-incrementalaccuracyNo clear effect
use incremental learning with prototypes instead of imbalanced direct trainingFLAR · SUN360 class-incrementalaccuracy at t=5Improved

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