【ECCV 2018】DeepJDOT: Deep Joint Distribution Optimal Transport for Unsupervised Domain Adaptation
0 likes424 views
1. The document discusses a convolutional neural network model for image classification. 2. The model uses convolutional layers to extract features from input images, followed by fully connected layers to perform classification. 3. Equations are provided to define the operations of the convolutional and fully connected layers in the neural network model.
【ECCV 2018】DeepJDOT: Deep Joint Distribution Optimal Transport for Unsupervised Domain Adaptation
- 1. ,. C BCA DC , C 1 .A1 B AC A / BD A B 1 1 C1C 1A1C DB 1 1 1A1 :1 A A 1 1A B .D 1 1B DAC JO N T R K CC A 1CBD 1C1 1 C A : C F 1 A 1 C
- 2. O 2 J e 2 / ) / g 2 J A D J 2c kT a 2
- 3. T 3 O !"($%, $') = *+,-∈/(01, 02) ∫42 5 6%, 6' 7Υ(6%, 6') O
- 4. 4 4 ) ( ) 4 O L O 4 D T J !(#$ % , '$ % , #( ) , '( ) ) = α- #( % , #( ) + /('( % , '( ) ) : ) ) 4 4 4 , ) ,
- 5. Source CNN Source Target Output CNN CNN
- 6. CNN !"
- 7. 7 7 ,
- 8. - -
- 10. 01