AI Paper Atlas

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

Universal Representation Learning From Multiple Domains for Few-Shot Classification

Wei-Hong Li, Xialei Liu, Hakan Bilen

ICCV 2021 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Embedding adaptation for few-shot classification · 107 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 CKA+KL distillation instead of L2 feature loss (ablation)URL · Meta-Dataset ImageNetaccuracy: 55.7 -> 58.8Improved
add CKA feature loss to KL (ablation)URL · Meta-Dataset Texturesaccuracy: 73.3 -> 77.2Improved
add CKA feature loss to KL (ablation)URL · Meta-Dataset MNISTaccuracy: 95.3 -> 94.7Hurt
add CKA feature loss to KL (ablation)URL · Meta-Dataset VGG Floweraccuracy: 89.6 -> 92.0Improved
use CKA instead of cosine distance (ablation)URL · Meta-Dataset VGG Floweraccuracy: 85.0 -> 91.3Improved
add distillation from domain-specific networks versus vanilla multi-domain learning (ablation)URL · Meta-Dataset Fungiaccuracy: 61.9 -> 68.1Improved
add distillation versus vanilla multi-domain learning (ablation)URL · Meta-Dataset Traffic Signaccuracy: 51.0 -> 63.3Improved
add linear feature adaptation versus nearest-centroid only (ablation)URL · Meta-Dataset Fungiaccuracy: 66.2 -> 68.1Improved
add linear feature adaptation versus Mahalanobis distance (ablation)URL · Meta-Dataset ImageNetaccuracy: 53.9 -> 58.8Improved
add learned adaptation versus SVM on Traffic Sign (ablation)URL · Meta-Dataset Traffic Signaccuracy: 65.7 -> 63.3Hurt
add feature adaptors A_theta in distillation (ablation)URL (CKA) · Meta-Dataset Birdsaccuracy: +1.7Improved
add feature adaptors A_theta in distillation (ablation)URL (CKA) · Meta-Dataset VGG Floweraccuracy: +3.6Improved

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