AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

Coarsely-Labeled Data for Better Few-Shot Transfer

Cheng Perng Phoo, Bharath Hariharan

ICCV 2021 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Fine-tuning strategies for few-shot · 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
add parent-consistent filtering of pseudo-labels (ablation)self-training student · iNat2019-CL pseudo-label groupingAMI: 0.4258 -> 0.7260Improved
add parent-consistent filtering of pseudo-labels (ablation)self-training student · tieredImageNet-CL pseudo-label groupingAMI: 0.4620 -> 0.7352Improved
add parent-consistent filtering of pseudo-labels (ablation)self-training student · CIFAR-100-CL pseudo-label groupingAMI: 0.3695 -> 0.8293Improved
add parent-consistent filtering (PAS vs Self-training) (ablation)PAS · iNat2019-CL all-way 5-shottop-1 per-class accuracy: 42.17 -> 43.27Improved
add parent-consistent filtering (PAS vs Self-training) (ablation)PAS · CIFAR-100-CL all-way 5-shottop-1 per-class accuracy: 42.43 -> 48.42Improved
use PAS instead of multitask training for coarse labelsrepresentation learning · iNat2019-CL all-way 5-shottop-1 per-class accuracy: 41.42 -> 43.27Improved
use coarsely-labeled data with PAS instead of baselinerepresentation learning · all-way 5-shot, three datasets averagetop-1 per-class accuracy: 8.31 points gain over BaselineImproved
train only on coarse supercategory labels (Repr-Coarse) instead of fine labelsrepresentation learning · tieredImageNet-CL all-way 5-shottop-1 per-class accuracy: 53.36 -> 37.19Hurt
reduce base classes from 398 to 144 (Baseline) (ablation)Baseline · iNat2019-CL all-way 5-shottop-1 per-class accuracy: 39.22 -> 28.37Hurt
reduce base classes from 398 to 144 (PAS) (ablation)PAS · iNat2019-CL all-way 5-shottop-1 per-class accuracy: 43.27 -> 37.31Hurt
add PAS with unfiltered pseudo-labels on unseen supercategoriesrepresentation learning · iNat2019-CL (unseen supercategories) all-way 5-shottop-1 per-class accuracy: 35.72 -> 40.73Improved
use coarser labels (kingdom or order) in PAS filtering instead of genus (ablation)PAS · iNat2019-CLtop-1 per-class accuracy: AMI 0.73 -> 0.64 -> 0.45Hurt

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