AI Paper Atlas

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

Spatial Ensemble: a Novel Model Smoothing Mechanism for Student-Teacher Framework

Tengteng Huang, Yifan Sun, Xun Wang, Haotian Yao, Chi Zhang

NeurIPS 2021 · Label-efficient and robust learning › Knowledge distillation · approach: Teacher properties and robustness · 5 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 spatial ensemble as the model smoothing instead of no smoothing (ablation)MoCo and BYOL · ImageNet linear evaltop-1 accuracy: fail -> 66.7 (MoCo), fail -> 67.8 (BYOL)Improved
use spatial-temporal smoothing instead of temporal moving averageMoCo · ImageNet linear evaltop-1 accuracy: 67.5 -> 67.8Improved
use spatial-temporal smoothing instead of temporal moving averageBYOL · ImageNet linear evaltop-1 accuracy: 71.8 -> 72.7Improved
use neuron-wise instead of layer-wise spatial ensemble (ablation)SE only (BYOL) · ImageNet linear evaltop-1 accuracy: 67.8 -> 71.6Improved
use neuron-wise instead of layer-wise granularity inside STS (ablation)STS (BYOL) · ImageNet linear evaltop-1 accuracy: 72.7 -> 72.8No clear effect
use neuron-wise instead of layer-wise granularity (ablation)STS (BYOL) · ImageNet linear evalrelative training time: 0.99x -> 1.82xHurt
use STS instead of TMA under ImageNet-C corruptionsBYOL · ImageNet-Cmean top-1 accuracy: around 1.5%Improved
add STS to the FixMatch teacherFixMatch · CIFAR-10 20 labelstop-1 accuracy: 86.68 -> 93.31Improved
add STS to the FixMatch teacherFixMatch · CIFAR-10 40 labelstop-1 accuracy: 92.06 -> 93.89Improved
add STS to the FixMatch teacherFixMatch · CIFAR-10 250 labelstop-1 accuracy: 94.89 -> 95.10Improved
add STS to the FixMatch teacherFixMatch · CIFAR-10 4000 labelstop-1 accuracy: 95.77 -> 95.94Improved
add STS to the FixMatch teacherFixMatch · CIFAR-100 200 labelstop-1 accuracy: 42.78 -> 45.85Improved

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