AI Paper Atlas

Research map / Continual learning / Class-incremental

Beyond Point-wise Neural Collapse: A Topology-Aware Hierarchical Classifier for Class-Incremental Learning

HuiYu Yi, Xu Zhiming, Dunwei Tu, Zhicheng Wang, Baile Xu, Furao Shen

ICML 2026 · Continual learning › Class-incremental · approach: Exemplar-efficient class-incremental learning · 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 original classifier with HC-SOINN in seven PTM-based CIL methodsSEMA · Split CUB-200 (20 tasks)last-task accuracy (%): 65.56 -> 86.56Improved
replace FC classifier with HC-SOINNCODA-Prompt · Split CUB-200 (20 tasks)last-task accuracy (%): 69.97 -> 85.75Improved
replace classifier with HC-SOINNCL-LoRA · Split CUB-200 (20 tasks)last-task accuracy (%): 74.47 -> 74.60No clear effect
replace classifier with HC-SOINNEASE · Split CUB-200 (20 tasks)last-task accuracy (%): 81.21 -> 81.13No clear effect
replace classifier with HC-SOINNCODA-Prompt · Split CIFAR-100 (10 tasks)last-task accuracy (%): 86.96 -> 87.56Improved
replace classifier with HC-SOINNDualPrompt · Split ImageNet-R (40 tasks)last-task accuracy (%): 58.67 -> 69.27Improved
add STAR drift alignment to HC-SOINN (ablation)DualPrompt + HC-SOINN · Split CUB-200 (20 tasks)last-task accuracy (%): 77.31 -> 85.84Improved
add STAR drift alignment to HC-SOINN (ablation)DualPrompt + HC-SOINN · Split ImageNet-R (40 tasks)last-task accuracy (%): 69.27 -> 69.75No clear effect
add STAR drift alignment to HC-SOINN (ablation)CODA-Prompt + HC-SOINN · Split CIFAR-100 (10 tasks)last-task accuracy (%): 87.56 -> 89.67Improved
use HC-SOINN instead of KAC classifierDualPrompt · Split ImageNet-R (40 tasks)last-task accuracy (%): 63.85 -> 69.75 (with STAR)Improved
KAC classifier instead of FC classifierDualPrompt · Split CIFAR-100 (10 tasks)average accuracy (%): 86.40 -> 86.30Hurt
combine hierarchical clustering with SOINN refinement versus either alone (ablation)HC-SOINN · Split CUB-200 (CODA-Prompt)last-task accuracy (%): 82.65 / 83.80 -> 85.75Improved

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