AI Paper Atlas

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

Overcoming Concept Shift in Domain-Aware Settings through Consolidated Internal Distributions

Mohammad Rostami, Aram Galstyan

AAAI 2023 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Conditional distribution alignment · 16 citations (OpenAlex)

Read the paper

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
Adapt the source-trained model using a GMM of its internal source distribution and sliced Wasserstein alignmentSMAUI · MNIST→USPSclassification accuracy: 92.2 $\pm$ 0.5 vs. \textbf{96.4}$\pm$ 0.3Hurt
Adapt the source-trained model using a GMM of its internal source distribution and sliced Wasserstein alignmentSMAUI · USPS→MNISTclassification accuracy: \textbf{98.2} $\pm$ 0.2 vs. 98.0Improved
Adapt the source-trained model using a GMM of its internal source distribution and sliced Wasserstein alignmentSMAUI · SVHN→MNISTclassification accuracy: 92.6 $\pm$ 1.0 vs. \textbf{97.9}$\pm$ 0.4Hurt
Adapt the source-trained model using internal-distribution alignmentSMAUI · Office-31 averageclassification accuracy: 88.4 vs. \textbf{90.6}Hurt
Adapt the source-trained model using internal-distribution alignmentSMAUI · ImageCLEF-DA averageclassification accuracy: \textbf{95.3} vs. 89.7Improved
Adapt the source-trained model using internal-distribution alignmentSMAUI · Office-Caltech averageclassification accuracy: \textbf{94.0} vs. 94.6Hurt
Adapt the source-trained model using internal-distribution alignmentSMAUI · VisDA-2017 Syn.→Real.classification accuracy: \textbf{76.9} $\pm$ 0.7 vs. 71.9Improved
Use high-confidence pseudo-samples when generating the internal-distribution datasetSMAUI · Sequential model adaptationtarget-domain generalizationMixed

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