AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Learning a Mixture of Granularity-Specific Experts for Fine-Grained Categorization

Lianbo Zhang, Shaoli Huang, Wei Liu, Dacheng Tao

ICCV 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 220 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 KL-divergence constraint between experts (ablation)two-expert MGE-CNN (ResNet-50) · CUB-200-2011top-1 accuracy: 87.9 -> 88.2Improved
add a second expert (ablation)MGE-CNN (ResNet-50) · CUB-200-2011top-1 accuracy: 86.8 -> 88.2Improved
add a third expert (ablation)MGE-CNN (ResNet-50) · CUB-200-2011top-1 accuracy: 88.2 -> 88.5Improved
combine second expert with first (ensemble) (ablation)MGE-CNN (ResNet-50) · CUB-200-2011top-1 accuracy: 87.3 -> 87.9Improved
add GMP small-part branch to GAP branch (concat) (ablation)first expert (ResNet-50) · CUB-200-2011top-1 accuracy: 85.4 -> 86.8Improved
add GMP small-part branch to GAP branch (concat) (ablation)third expert (ResNet-50) · CUB-200-2011top-1 accuracy: 85.2 -> 86.1Improved
use GMP branch alone instead of GAP (ablation)first expert (ResNet-50) · CUB-200-2011top-1 accuracy: 85.4 -> 83.8Hurt
raise attention threshold from 0.2 to 0.3 (ablation)two-expert MGE-CNN · CUB-200-2011top-1 accuracy: 88.19 -> 88.44No clear effect
use ResNet-101 backbone instead of ResNet-50 (ablation)MGE-CNN (3 experts) · CUB-200-2011top-1 accuracy: 88.5 -> 89.4Improved
use ResNet-101 backbone instead of ResNet-50 (ablation)MGE-CNN · Stanford Carstop-1 accuracy: 93.9 -> 93.6No clear effect
add MGE-CNN to ResNet-50ResNet-50 · CUB-200-2011top-1 accuracy: 85.4 -> 88.5Improved
add MGE-CNN to ResNet-50ResNet-50 · Stanford Carstop-1 accuracy: 91.7 -> 93.9Improved

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