AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

Learning Compatible Embeddings

Qiang Meng, Chixiang Zhang, Xiaoqiang Xu, Feng Zhou

ICCV 2021 · Label-efficient and robust learning › Knowledge distillation · approach: Few-shot and metric distillation · 38 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 boundary loss to alignment loss (ablation)LCE (alignment only, no transformation) · IJB-C (ResNet50, MS1Mv2)upgrade gain (%): 60.82 -> 68.34Improved
add boundary loss to alignment loss (ablation)LCE (alignment only, no transformation) · IJB-C (ResNet50, MS1Mv2)performance gain (%): 84.28 -> 89.07Improved
add transformation module to alignment loss (ablation)LCE (alignment only) · IJB-C (ResNet50, MS1Mv2)performance gain (%): 84.28 -> 96.36Improved
add transformation module to alignment plus boundary loss (ablation)LCE (alignment+boundary) · IJB-C (ResNet50, MS1Mv2)performance gain (%): 89.07 -> 99.32Improved
add transformation module to alignment loss (ablation)LCE (alignment only) · IJB-C (ResNet50, MS1Mv2)upgrade gain (%): 60.82 -> 67.43Improved
add boundary loss to alignment plus transformation (ablation)LCE (alignment+transformation) · IJB-C (ResNet50, MS1Mv2)upgrade gain (%): 67.43 -> 72.44Improved
add transformation module to LCE (direct) (ablation)LCE with ResNet18 new model · IJB-C (ResNet18, old ResNet50)performance gain (%): -114.91 -> -95.63Mixed
use LCE point-to-set constraints instead of BCTcompatible training · IJB-C (training data change)upgrade gain (%): 22.54 -> 81.37Improved
use LCE instead of BCTcompatible training · IJB-C (new backbone ResNet100)upgrade gain (%): 52.04 -> 90.82Improved
use LCE instead of BCTcompatible training · IJB-C (new backbone MobileFace)upgrade gain (%): -45.33 -> -47.50No clear effect
use LCE instead of BCTcompatible training · IJB-C (new loss CosFace)upgrade gain (%): -75.38 -> -47.69Improved
use LCE instead of RBT when transformation trained with NormFacecross model compatibility · IJB-C (phi*_a, phi*_b)cross verification TAR@FAR=1e-4 (%): 92.78 -> 94.08Improved

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