Impact of ImageNet Model Selection on Domain Adaptation

In this paper, we aim to explore the effect of different ImageNet models on domain adaption, and want to find how the features from these deep neural networks affect the final domain transfer accuracy.

We are the first to examine how different ImageNet models affect domain transfer accuracy, using features from sixteen distinct pre-trained neural networks on twelve methods across three benchmark datasets. The correlation of domain adaptation performance and ImageNet classification performance is high, ranging from 0.71 to 0.95, suggesting that features from a higher-performing ImageNet-trained model are more valuable than those from a lower-performing model.

We also find that all three benchmark datasets suggest that the layer prior to the last fully connected layer is the best source

T-SNE view of extracted features from the last fully connected layer of sixteen neural networks. Different colors represent different classes. The more separation of the classes in the dataset, the better the features are (from Amazon domain in the Office31 dataset).