… a novel regularized transfer learning framework DELTA, namely DEep Learning Transfer
using Feature Map with Attention. Instead of constraining the weights of neural network, DELTA …

… use attention maps generated from feature maps or Jacobians for transferring the source …
Our goal is learning to learn transfer rules for performing knowledge transfer in an automatic …

… a deep learning approach to learn a feature mapping … a deep learning method to
simultaneously learn a feature transfor… from lower-level features to higher-level features in the …

… Feature transfer aims to identify a joint feature map shared by the source and target domains…
domains have shared knowledge and representations, features transferred from the source …

… In contrast, traditional machine learning methods need to design features manually that …
Mapping-based deep transfer learning refers to mapping instances from the source domain and …

… transfer learning is successful in learning features relevant not only for nighttime light prediction
but also for poverty mapping. … these features are highly informative for poverty mapping …

… to feature learning for transfer learning are required. In this section, we first introduce a
feature learning approach to transfer learning … Another context is to learn a feature mapping to …

… The number of feature maps, filter size, and size of the feature maps are hyperparameters in
… that one has to optimize for a learning task. A deep CNN architecture is multi-layered, with …

… Meanwhile, we introduce feature map loss and map … feature map loss produced by
discriminator is beneficial for computing the discriminative regions from both real and generated …

… The characteristic kernel associated with the feature map φ, k (xs, xt) = <φ (xs) ,φ (xt)>, is …
We compare the DAN model to state-of-the-art transfer learning and deep learning methods on …