AI Paper Atlas

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

Creativity Inspired Zero-Shot Learning

Mohamed Elhoseiny, Mohamed Elfeki

ICCV 2019 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Generative feature synthesis for zero-shot · 67 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 entropy-maximization term on hallucinated text (ablation)GAZSL+CIZSL · CUB easy splitSeen-Unseen AUC: 33.3 -> 39.2Improved
add entropy-maximization term on hallucinated text (ablation)GAZSL+CIZSL · CUB hard splitSeen-Unseen AUC: 10.1 -> 11.9Improved
add realism (discriminator real) term on hallucinated text (ablation)GAZSL+CIZSL · CUB easy splitSeen-Unseen AUC: 35.0 -> 39.2Improved
add realism (discriminator real) term on hallucinated text (ablation)GAZSL+CIZSL · CUB hard splitSeen-Unseen AUC: 8.2 -> 11.9Improved
use entropy deviation instead of classifying hallucinated text as a new class (ablation)GAZSL+CIZSL · CUB hard splitSeen-Unseen AUC: 9.5 -> 11.9Improved
learn Sharma-Mittal entropy parameters instead of fixed KL (ablation)GAZSL+CIZSL · CUB hard splitSeen-Unseen AUC: 11.6 -> 11.9Improved
add CIZSL loss to GAZSLGAZSL · CUB easy splitSeen-Unseen AUC: 35.4 -> 39.2Improved
add CIZSL loss to GAZSLGAZSL · CUB hard splitSeen-Unseen AUC: 8.7 -> 11.9Improved
add CIZSL loss to FeatGenFeatGen · CUB easy splitSeen-Unseen AUC: 34.1 -> 37.4Improved
add CIZSL loss to GAZSLGAZSL · AwA2 attribute ZSLTop-1 accuracy: 58.9 -> 67.8Improved
add CIZSL loss to FeatGenFeatGen · SUN attribute ZSLTop-1 accuracy: 60.8 -> 59.4Hurt
add CIZSL loss to GAZSLGAZSL · CUB zero-shot retrieval (100%)mAP: 40.3 -> 46.2Improved

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