AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

S2-Boost: Synergistic Semantic Boosting for Coarse-to-Fine Ensemble Learning

Guanxiong He, Zheng Wang, Jie Wang, Liaoyuan Tang, Rong Wang, Feiping Nie

AAAI 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded visual concept learning · 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 boosting over semantic partitions (ablation)S2-Boosting · ImageNet-1KTop-1 accuracy: 76.73 -> 88.35Improved
use learned semantic masks instead of random grid masks (ablation)S2-Boosting · ImageNet-1KTop-1 accuracy: 85.07 -> 88.35Improved
add diversity and center losses to SPSLM (ablation)S2-Boosting · ImageNet-1KTop-1 accuracy: 83.35 -> 88.35Improved
add reconstruction loss to SBM (ablation)S2-Boosting · ImageNet-1KTop-1 accuracy: 81.37 -> 88.35Improved
order masks by mean activation (coarse-to-fine) instead of random order (ablation)SBM · ImageNet-1KTop-1 accuracyImproved
use S2-Boosting instead of plain ViT-LViT-L · ImageNet-1KTop-1 accuracy: 85.15 -> 88.35Improved
use S2-Boosting instead of Re-BoostingViT-L · ImageNet-1KTop-1 accuracy: 84.71 -> 88.35Improved
use S2-Boosting instead of Re-BoostingViT-L · CIFAR-10Top-1 accuracy: 99.24 -> 99.51Improved
use S2-Boosting instead of MAEViT-L · CIFAR-100Top-1 accuracy: 93.31 -> 93.47Improved
apply S2-Boosting under 50% random patch occlusionViT-L · ImageNet-1K (50% occlusion)Top-1 accuracy: 60.37 -> 75.61Improved
apply S2-Boosting under 20% random patch occlusionViT-L · ImageNet-1K (20% occlusion)Top-1 accuracy: 77.88 -> 82.19Improved

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