AI Paper Atlas

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

AlignFlow: Cycle Consistent Learning from Multiple Domains via Normalizing Flows

Aditya Grover, Christopher Chute, Rui Shu, Zhangjie Cao, Stefano Ermon

AAAI 2020 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Conditional distribution alignment · 49 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
combine adversarial and MLE trainingAlignFlow · FacadesMSE: 0.7129 -> 0.5801; 0.3286 -> 0.2512Improved
combine adversarial and MLE trainingAlignFlow · MapsMSE: 0.0245 -> 0.0209; 0.0953 -> 0.0897Improved
combine adversarial and MLE trainingAlignFlow · CityScapesMSE: 0.1252 -> 0.1130; 0.1200 -> 0.1462Mixed
add the MLE objective to adversarial training (ablation)AlignFlow · FacadesMSE: 0.6727 -> 0.5801; 0.2679 -> 0.2512Improved
add adversarial training to MLE training (ablation)AlignFlow · FacadesMSE: 0.9014 -> 0.5801; 0.5960 -> 0.2512Improved
add the MLE objective to adversarial training (ablation)AlignFlow · MapsMSE: 0.0385 -> 0.0209; 0.1123 -> 0.0897Improved
add adversarial training to MLE training (ablation)AlignFlow · MapsMSE: 0.0452 -> 0.0209; 0.1746 -> 0.0897Improved
add the MLE objective to adversarial training (ablation)AlignFlow · CityScapesMSE: 0.2569 -> 0.1130; 0.2196 -> 0.1462Improved
add adversarial training to MLE training (ablation)AlignFlow · CityScapesMSE: 0.2526 -> 0.1130; 0.2272 -> 0.1462Improved
replace CycleGAN translation with AlignFlowCyCADA adaptation pipeline · MNIST→USPSclassification accuracy: 95.6 ± 0.2 -> 96.2 ± 0.2Improved
replace CycleGAN translation with AlignFlowCyCADA adaptation pipeline · USPS→MNISTclassification accuracy: 96.5 ± 0.1 -> 96.7 ± 0.1Improved
replace CycleGAN translation with AlignFlowCyCADA adaptation pipeline · SVHN→MNISTclassification accuracy: 90.4 ± 0.4 -> 91.0 ± 0.3Improved

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