AI Paper Atlas

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

Improving Knowledge Distillation via Category Structure

Zailiang Chen, Xianxian Zheng, Hailan Shen, Ziyang Zeng, Yukun Zhou, Rongchang Zhao

ECCV 2020 · Label-efficient and robust learning › Knowledge distillation · approach: Distillation with causal and adaptive guidance · 21 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 intra-category structure loss to KD (ablation)KD student · Tiny ImageNettop-1 accuracy: 49.53 -> 52.51Improved
add inter-category structure loss to KD (ablation)KD student · Tiny ImageNettop-1 accuracy: 49.53 -> 52.48Improved
add inter-category loss on top of intra-category loss (ablation)CSKD · Tiny ImageNettop-1 accuracy: 52.51 -> 53.66Improved
add intra-category loss on top of inter-category loss (ablation)CSKD · Tiny ImageNettop-1 accuracy: 52.48 -> 53.66Improved
add inter-category loss on top of intra-category loss (ablation)CSKD · Tiny ImageNettop-5 accuracy: 78.36 -> 78.75Improved
add both category structure losses to KD (ablation)KD student · Tiny ImageNettop-5 accuracy: 74.90 -> 78.75Improved
use CSKD instead of CCKDstudent ResNet18 0.25 · Tiny ImageNettop-1 accuracy: 53.14 -> 53.66Improved
use CSKD instead of KDstudent ResNet18 0.25 · CIFAR-10accuracy: average 1.52% improvementImproved
use CSKD instead of CE-only trainingstudent networks · CIFAR-100accuracy: 2.82%~4.15%Improved
distill with CSKD from R101 0.5 teacher into R50 0.25student ResNet50 0.25 · CIFAR-10accuracy: 92.91 -> 93.20Improved
distill with CSKD from R152 0.5 teacher into R50 0.25student ResNet50 0.25 · CIFAR-100accuracy: 72.15 -> 72.60Improved
use CSKD instead of KDGAN (R101 0.5/R18 0.25)student ResNet18 0.25 · CIFAR-10accuracy: 91.54 -> 91.99Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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