Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Deep Domain Confusion: Maximizing for Domain Invariance

About

Recent reports suggest that a generic supervised deep CNN model trained on a large-scale dataset reduces, but does not remove, dataset bias on a standard benchmark. Fine-tuning deep models in a new domain can require a significant amount of data, which for many applications is simply not available. We propose a new CNN architecture which introduces an adaptation layer and an additional domain confusion loss, to learn a representation that is both semantically meaningful and domain invariant. We additionally show that a domain confusion metric can be used for model selection to determine the dimension of an adaptation layer and the best position for the layer in the CNN architecture. Our proposed adaptation method offers empirical performance which exceeds previously published results on a standard benchmark visual domain adaptation task.

Eric Tzeng, Judy Hoffman, Ning Zhang, Kate Saenko, Trevor Darrell• 2014

Related benchmarks

TaskDatasetResultRank
3D Human Pose EstimationMPI-INF-3DHP (test)--
559
3D Human Pose Estimation3DPW (test)
PA-MPJPE75.3
505
Image ClassificationOffice-31
Average Accuracy86.9
261
Domain AdaptationOffice-31 unsupervised adaptation standard
Accuracy (A to W)75.8
162
Domain AdaptationOffice-31
Accuracy (A -> W)61
156
Domain AdaptationOFFICE
Average Accuracy78.3
96
Image ClassificationOffice-31 (test)
Avg Accuracy70.7
93
Unsupervised Domain AdaptationOffice-31
A->W Accuracy75.6
83
Image ClassificationOffice-10 + Caltech-10
Average Accuracy88.2
77
Image ClassificationSVHN to MNIST (test)
Accuracy71.1
66
Showing 10 of 44 rows

Other info

Follow for update