AI Paper Atlas

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

Beyond Student: An Asymmetric Network for Neural Network Inheritance

Yiyun Zhou, Jingwei Shi, Mingjing Xu, Zhonghua Jiang, Jingyuan Chen

ICLR 2026 · Label-efficient and robust learning › Knowledge distillation · approach: Distillation with causal and adaptive guidance · 1 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 SVD initialization (ablation)InherNet Large · CIFAR-100 (ResNet56 -> ResNet20)top-1 accuracy (%): 69.35 -> 73.67Improved
add SVD initialization (ablation)InherNet Large · CIFAR-100 (ResNet32x4 -> ResNet8x4)top-1 accuracy (%): 76.68 -> 78.53Improved
add gating mechanism (ablation)InherNet Large · CIFAR-100 (ResNet56 -> ResNet20)top-1 accuracy (%): 73.22 -> 73.67Improved
use asymmetric instead of symmetric structure (ablation)InherNet Large · CIFAR-100 (ResNet56 -> ResNet20)top-1 accuracy (%): 73.28 -> 73.67Improved
use one-down-many-ups instead of inverted many-downs-one-up (ablation)InherNet Large · CIFAR-100 (VGG13 -> VGG8)top-1 accuracy (%): 71.26 -> 75.16Improved
use one-down-many-ups instead of inverted many-downs-one-up (ablation)InherNet Large · CIFAR-100 (ResNet32x4 -> ResNet8x4)top-1 accuracy (%): 78.79 -> 78.53Hurt
use one-down-many-ups instead of inverted many-downs-one-up (ablation)InherNet · GLUE MRPCaccuracy (%): 87.36 -> 88.74Improved
replace KD student with InherNet LargeResNet20 student · CIFAR-100 (ResNet56 teacher)top-1 accuracy (%): 72.33 -> 73.67Improved
replace KD student with InherNet LargeResNet8x4 student · CIFAR-100 (ResNet32x4 teacher)top-1 accuracy (%): 78.28 -> 78.53Improved
replace T5-Small+KD with InherNetT5-Small · GLUE (8 tasks)accuracy/MCC (%): MNLI 53.81 -> 56.34; RTE 74.31 -> 72.93; QQP 84.92 -> 84.69Mixed
replace CLIP-KD student with InherNetResNet-101 teacher · ImageNet zero-shot (CC3M pretrain)top-1 accuracy (%): 17.30 -> 17.39Improved
replace CLIP-KD EfficientNet-B0 with InherNetResNet-101 teacher · CC3M I2TR@1 (%): 33.65 -> 32.17Hurt

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