AI Paper Atlas

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

Predicting Visual Exemplars of Unseen Classes for Zero-Shot Learning

Soravit Changpinyo, Wei-Lun Chao, Fei Sha

ICCV 2017 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Generative feature synthesis for zero-shot · 176 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
predict visual exemplars and use as semantic representations instead of raw attributesSynC struct · AwAaccuracy (%): 72.9 -> 77.2Improved
use predicted exemplars with ConSEConSE · CUBaccuracy (%): 36.2 -> 46.2Improved
use predicted exemplars with LatEmLatEm · CUBaccuracy (%): 48.0 -> 56.2Improved
use predicted exemplars with ConSEConSE · AwAaccuracy (%): 63.3 -> 70.5Improved
use predicted exemplars with SynC structSynC struct · SUNaccuracy (%): 62.7 -> 66.1Improved
use 1NN on predicted exemplars instead of SynC o-vs-oSynC o-vs-o · ImageNet 2-hopFlat Hit@1 (%): 10.5 -> 11.7Improved
use standardized Euclidean distance (1NNs) instead of Euclidean (1NN) (ablation)EXEM (1NN) · CUBaccuracy (%): 56.3 -> 58.5Improved
use standardized Euclidean distance (1NNs) instead of Euclidean (1NN) (ablation)EXEM (1NN) · SUNaccuracy (%): 69.6 -> 67.3Hurt
add PCA projection d=1024 before regression (ablation)EXEM (1NN) · AwAaccuracy (%): 77.8 -> 76.2Hurt
add PCA projection d=1024 before regression (ablation)EXEM (1NN) · CUBaccuracy (%): 55.1 -> 56.3Improved
reduce PCA dimension from 1024 to 500 (ablation)EXEM (1NN) · AwAaccuracy (%): 76.2 -> 76.2No clear effect
use SVR instead of 2-layer MLP for exemplar prediction (ablation)EXEM (1NN) · CUB split 1accuracy (%): 53.8 -> 57.1Improved

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