AI Paper Atlas

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

Distilling Balanced Knowledge from a Biased Teacher

Seonghak Kim

CVPR 2026 · Label-efficient and robust learning › Knowledge distillation · approach: Rethinking distillation objectives

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 LTKD instead of standard KD/DKDResNet32x4 -> ResNet8x4 · CIFAR-100-LT (gamma=100)tail accuracy: 15.09 -> 27.21Improved
add LTKDResNet50 -> MobileNetV2 · CIFAR-100-LT (gamma=100)tail and overall accuracy: +8.59 tail, +3.24 overallImproved
add LTKDResNet32x4 -> ResNet8x4 · TinyImageNet-LT (gamma=100)tail accuracy: 9.09 -> 10.48Improved
add LTKDResNet34 -> ResNet18 · ImageNet-LTtail accuracy: +1.06, +1.75, +2.40 for gamma 5, 10, 20Improved
add rebalanced cross-group loss (ablation)LTKD (cross-group only) · CIFAR-100-LT (gamma=20), ResNet32x4-ResNet8x4tail accuracy: 38.30 -> 40.51Improved
add rebalanced cross-group loss (ablation)LTKD (cross-group only) · CIFAR-100-LT (gamma=20), ResNet32x4-ShuffleNetV1overall accuracy: -0.01 overall, tail 30.01 -> 30.68No clear effect
add cross-group loss alone (ablation)LTKD · CIFAR-100-LT (gamma=20), ResNet32x4-ResNet8x4tail accuracy: +3.70Improved
add cross-group loss alone (ablation)LTKD · CIFAR-100-LT (gamma=20), ResNet32x4-ResNet8x4overall accuracy: 57.41 -> 56.55Hurt
add reweighted within-group loss alone (ablation)LTKD · CIFAR-100-LT (gamma=20), ResNet32x4-ResNet8x4tail accuracy: +5.53Improved
combine cross- and within-group losses (ablation)LTKD · CIFAR-100-LT (gamma=20), ResNet32x4-ResNet8x4tail accuracy: 42.34 -> 49.70Improved
combine both losses (ablation)LTKD · CIFAR-100-LT (gamma=20), VGG13-MobileNetV2tail and overall accuracy: +9.28 tail, +4.34 overallImproved
increase number of groups from 3 to 10-20 (ablation)LTKD · CIFAR-100-LT (gamma=100)overall accuracy: VGG13-VGG8 47.66 -> 48.69 (n=20); R32x4-R8x4 51.08 -> 50.41 (n=100)Mixed

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