AI Paper Atlas

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

Towards Learning Spatially Discriminative Feature Representations

Chaofei Wang, Jiayu Xiao, Yizeng Han, Qisen Yang, Shiji Song, Gao Huang

ICCV 2021 · Label-efficient and robust learning › Knowledge distillation · approach: Few-shot and metric distillation · 25 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 CAM-lossResNet-56 · CIFAR-100top-1 error (%): 28.80 -> 27.34Improved
add CAM-lossResNet-50 · ImageNet-1Ktop-1 error (%): 23.68 -> 22.98Improved
add CAM-lossResNext-101 8x32d · ImageNet-1Ktop-1 error (%): 21.04 -> 20.45Improved
add CAM-lossDenseNet-bc-190-40 · CIFAR-100top-1 error (%): 17.67 -> 16.98Improved
add CAM-loss to ShakeDropPyramidNet-200 · CIFAR-100top-1 error (%): 15.08 -> 14.56Improved
add CAM-loss to CutMixPyramidNet-200 · CIFAR-100top-1 error (%): 14.47 -> 14.01Improved
add CAM-loss to CutMix + ShakeDropPyramidNet-200 · CIFAR-100top-1 error (%): 13.81 -> 13.49Improved
add CAM-loss to CutMixResNet-50 · ImageNet-1Ktop-1 error (%): 21.54 -> 21.16Improved
use CAM-loss instead of NegativeCAMResNet-50 · ImageNet-1Ktop-1 error (%): 23.32 -> 22.98Improved
use CAM-loss instead of NegativeCAMResNet-56 · CIFAR-100top-1 error (%): 27.37 -> 27.34No clear effect
use CAM-loss instead of center lossResNet-110 · CIFAR-100top-1 error (%): 26.88 -> 26.56Improved
pretrain with CAM-loss then fine-tuneResNet-50 · CUB transfertop-1 accuracy (%): 85.6 -> 86.7Improved

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