AI Paper Atlas

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

Evolving Standardization for Continual Domain Generalization over Temporal Drift

Mixue Xie, Shuang Li, Longhui Yuan, Chi Harold Liu, Zehui Dai

NeurIPS 2023 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Foundation-model domain generalization · 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 stop-gradient sg() in L_stan (ablation)EvoS · YearbookOOD avg accuracy (%): 55.37 -> 93.97Improved
add stop-gradient in L_ce and L_stan (vs only L_stan) (ablation)EvoS · YearbookOOD avg accuracy (%): 93.97 -> 95.53Improved
add truncation when sampling features (ablation)EvoS · YearbookOOD avg accuracy (%): 94.46 -> 95.53Improved
add adversarial and consistency losses (ablation)EvoS · YearbookOOD avg accuracy (%): 93.25 -> 95.53Improved
add consistency loss L_con (ablation)EvoS · YearbookOOD avg accuracy (%): 94.75 -> 95.53Improved
add adversarial loss L_adv (ablation)EvoS · YearbookOOD avg accuracy (%): 94.27 -> 95.53Improved
multi-scale attention instead of single-scale (ablation)EvoS · YearbookOOD avg accuracy (%): 94.09 -> 95.53Improved
use all historical domain distributions instead of one random (ablation)EvoS · YearbookOOD avg accuracy (%): 95.40 -> 95.53Improved
use all historical distributions instead of one random (ablation)EvoS · RMNISTOOD worst accuracy (%): 85.46 -> 87.04Improved
use all historical distributions instead of one random (ablation)EvoS · fMoWOOD worst accuracy (%): 44.76 -> 45.86Improved
use all historical distributions instead of one random (ablation)EvoS · HuffpostOOD avg accuracy (%): 71.57 -> 72.36Improved
use all historical distributions instead of one random (ablation)EvoS · ArxivOOD worst accuracy (%): 49.52 -> 50.19Improved

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