AI Paper Atlas

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

An Empirical Study and Analysis of Generalized Zero-Shot Learning for Object Recognition in the Wild

Wei-Lun Chao, Soravit Changpinyo, Boqing Gong, Fei Sha

ECCV 2016 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Generative feature synthesis for zero-shot · 614 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
use calibrated stacking instead of Gaussian novelty detectionSynC struct · AwAAUSUC: 0.424 -> 0.583Improved
use calibrated stacking instead of LoOP novelty detectionSynC struct · CUBAUSUC: 0.224 -> 0.356Improved
use calibrated stacking instead of Gaussian novelty detectionSynC struct · SUNAUSUC: 0.151 -> 0.260Improved
calibrate seen-class scores (gamma) instead of direct stackingSynC struct · AwAunseen-to-all accuracy (%): 0.4 -> 73.4Improved
select hyperparameters by AUSUC instead of accuracies (ablation)SynC o-vs-o · AwAAUSUC: 0.539 -> 0.568Improved
select hyperparameters by AUSUC instead of accuracies (ablation)ConSE · AwAAUSUC: 0.443 -> 0.428Hurt
select hyperparameters by AUSUC instead of accuracies (ablation)DAP · CUBAUSUC: 0.202 -> 0.194Hurt
select hyperparameters by AUSUC instead of accuracies (ablation)SynC struct · CUBAUSUC: 0.356 -> 0.356No clear effect
replace word2vec embeddings with G-attr from all images (ablation)SynC o-vs-o · ImageNet-2KAUSUC, Flat hit@1: 0.04 -> 0.25Improved
increase labeled images for G-attr from 1 to 100 (ablation)SynC o-vs-o · ImageNet-2KAUSUC, Flat hit@1: 0.08 -> 0.25Improved
replace human visual attributes with G-attr from all images (ablation)SynC o-vs-o · CUBAUSUC: 0.34 -> 0.54Improved
replace visual attributes with 1-shot G-attr (ablation)SynC o-vs-o · AwAAUSUC: 0.57 -> 0.55Hurt

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