AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

A Style and Semantic Memory Mechanism for Domain Generalization

Yang Chen, Yu Wang, Yingwei Pan, Ting Yao, Xinmei Tian, Tao Mei

ICCV 2021 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Optimization-based domain generalization · 45 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-domain style invariance loss and orthogonal regularization (ablation)Vanilla · Digits-DG averageaccuracy: 73.7 -> 77.9Improved
add intra-domain style invariance loss and orthogonal regularization (ablation)Vanilla · PACS averageaccuracy: 79.5 -> 83.0Improved
add inter-domain semantic invariance (jury) loss (ablation)Vanilla · Digits-DG averageaccuracy: 73.7 -> 80.5Improved
add inter-domain semantic invariance (jury) loss (ablation)Vanilla · PACS averageaccuracy: 79.5 -> 84.4Improved
add both style and semantic invariance losses (ablation)Vanilla · Digits-DG averageaccuracy: 73.7 -> 83.0Improved
add both style and semantic invariance losses (ablation)Vanilla · PACS averageaccuracy: 79.5 -> 86.6Improved
add semantic invariance loss on top of style invariance (ablation)Vanilla-style · PACS averageaccuracy: 83.0 -> 86.6Improved
add style invariance on top of semantic invariance (ablation)Vanilla-semantic · Digits-DG averageaccuracy: 80.5 -> 83.0Improved
replace style bank and instance contrastive loss with a domain classifier (ablation)STEAM · Design choice comparisonaccuracyHurt
replace jury semantic loss with L2-matching of same-class pairs (ablation)STEAM · Design choice comparisonaccuracyHurt
replace jury semantic loss with plain Info-NCE contrastive loss (ablation)STEAM · Design choice comparisonaccuracyHurt
use STEAM instead of VanillaVanilla · Office-Home averageaccuracy: 64.7 -> 66.8Improved

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