AI Paper Atlas

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

Z-Score Normalization, Hubness, and Few-Shot Learning

Nanyi Fei, Yizhao Gao, Zhiwu Lu, Tao Xiang

ICCV 2021 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Alignment-based few-shot metric learning · 226 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 z-score normalizationProtoNet · miniImageNet 1-shot (ResNet-12)5-way accuracy (%): 63.38 -> 66.06Improved
add z-score normalizationIMP · miniImageNet 1-shot (ResNet-12)5-way accuracy (%): 63.70 -> 65.01Improved
add z-score normalizationIEPT · miniImageNet 1-shot (ResNet-12)5-way accuracy (%): 64.05 -> 67.35Improved
add z-score normalizationProtoNet · miniImageNet 5-shot (ResNet-12)5-way accuracy (%): 81.22 -> 81.73Improved
add z-score normalizationProtoNet · tieredImageNet 1-shot (Conv4-64)5-way accuracy (%): 53.56 -> 56.70Improved
add z-score normalizationIEPT · CUB 1-shot (Conv4-64)5-way accuracy (%): 68.27 -> 73.54Improved
add z-score normalization to pre-trained features (ablation)pre-trained model · miniImageNet 1-shot5-way accuracy (%): 49.89 -> 52.80Improved
add z-score normalization to pre-trained features (ablation)pre-trained model · tieredImageNet 1-shot5-way accuracy (%): 51.12 -> 53.76Improved
add z-score normalization to pre-trained features (ablation)pre-trained model · miniImageNet 1-shot skewnessskewness of k-occurrence (lower is better): 3.219 -> 2.360Improved
use ZN instead of SimpleShot centering (ablation)pre-trained model · miniImageNet 1-shot5-way accuracy (%): 53.21 -> 52.80Hurt
use ZN instead of SimpleShot centering (ablation)pre-trained model · tieredImageNet 1-shot5-way accuracy (%): 52.89 -> 53.76Improved
use ZN instead of l2 normalization (ablation)ProtoNet episodic training · miniImageNet 1-shot5-way accuracy (%): 53.99 -> 55.16Improved

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