AI Paper Atlas

Research map / Continual learning / Class-incremental

PRISM: Progressive Robust Learning for Open-World Continual Category Discovery

Wei Feng, Sijin Zhou, Yiwen Jiang, Zongyuan Ge

ICLR 2026 · Continual learning › Class-incremental · approach: Class-balanced knowledge retention

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 PRISM instead of VB-CGCDcontinual category discovery · DomainNet Real->PaintingcACC (All): 57.3 -> 60.9 (Real); 32.4 -> 39.2 (Painting)Improved
use PRISM instead of VB-CGCDcontinual category discovery · CUB-CcACC (All): 34.2 -> 49.3 (clean); 31.7 -> 44.0 (corrupted)Improved
use PRISM instead of VB-CGCD on hard domainscontinual category discovery · DomainNet Real->Quickdraw (unknown domain)cACC (All): 6.0 -> 7.1Improved
add HCS (high-frequency component separation) and SAM (ablation)PRISM · DomainNet Real->PaintingcACC (All): 28.7 -> 35.0 (Painting)Improved
add IKT module on top of HCS (ablation)PRISM · DomainNet Real->PaintingcACC (All, Painting New): 32.5 -> 38.2 (New, Painting)Improved
add SAM on top of HCS+IKT (ablation)PRISM · DomainNet Real->PaintingcACC (All): 33.2 -> 39.2 (Painting)Improved
replace entropy/energy/raw-image separation with HCS (ablation)PRISM · DomainNet Real->PaintingcACC (All): 30.6 -> 39.2 (Painting, vs energy-based)Improved

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