Transferrable Prototypical Networks for Unsupervised Domain Adaptation

By Yingwei Pan, Ting Yao, Yehao Li, Yu Wang, Chong-Wah Ngo, Tao Mei. 2019

395 citations recorded by the source corpus.

In this paper, we introduce a new idea for unsupervised domain adaptation via a remold of Prototypical Networks, which learn an embedding space and perform classification via a remold of the distances to the prototype of each class. Specifically, we present Transferrable Prototypical Networks (TPN) for adaptation such that the prototypes for each class in source and target domains are close in the embedding space and the score distributions predicted by prototypes separately on source and target data are similar. Technically, TPN initially matches each target example to the nearest prototype in the source domain and assigns an example a ``pseudo" label. The prototype of each class could then be computed on source-only, target-only and source-target data, respectively. The optimization of TPN is end-to-end trained by jointly minimizing the distance across the prototypes on three types of data and KL-divergence of score distributions output by each pair of the prototypes. Extensive experiments are conducted on the transfers across MNIST, USPS and SVHN datasets, and superior results are reported when comparing to state-of-the-art approaches. More remarkably, we obtain an accuracy of 80.4% of single model on VisDA 2017 dataset.

Subfields: Artificial Intelligence, Cancer Research, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cancer Research

In Computer Vision and Pattern Recognition

Citations in the corpus

Prototype-Guided Continual Adaptation for Class-Incremental Unsupervised Domain Adaptation (2022, methodology): “However, a simple combination of existing methods [32,35,36] is not feasible for CI-UDA, since obtaining image prototypes for knowledge retaining [36] requires data labels but the target domain in CIUDA is totally unlabeled.”
Unsupervised Model Adaptation for Continual Semantic Segmentation (2020, methodology): “Recently, this property have been used for UDA (Pan et al. 2019; Chen et al. 2019a), where the means for distribution modes are considered as the class prototype.”
A Comprehensive Survey on Transfer Learning (2019, methodology): “[150] constructed three prototypical networks (corresponding toDS ,DT , and DS ∪ DT ) and incorporated the thought of multimodel consensus.”
Unsupervised Domain Adaptation via Structurally Regularized Deep Clustering (2020, background): “We note that quite a few recent UDA methods [13, 27, 41, 51] consider clustering of target data as well; however, they still do explicit feature alignment between the two domains via alignment of cluster centers/samples, thus prone to the aforementioned risk of damaged intrinsic target discriminatio”
Harmonizing Transferability and Discriminability for Adapting Object Detectors (2020, background): “, prototype alignment [58, 5, 37, 10] or entropy regularization [64, 47, 8]) would be no longer applicable.”
Sequential Unsupervised Domain Adaptation through Prototypical Distributions (2020, methodology): “The mean for each mode can be considered as the class prototype in the previous works [26, 27].”
Open Compound Domain Adaptation (2019, methodology): “Inspired by [45, 36, 28] on prototype analysis, we also use class centroids {ck}k=1 to construct our memoryM , where K is the number of object classes.”
Minimum Class Confusion for Versatile Domain Adaptation (2019, neutral): “TPN [29] is based on pseudo class-prototypes.”
Reliable Weighted Optimal Transport for Unsupervised Domain Adaptation (2020, background): “Inspired by the prototypical networks which has achieved signiﬁcant performance in domain adaptation [27], we exploit the prototypical information in the feature space to mitigate the wrong transport procedure.”
Stochastic Classifiers for Unsupervised Domain Adaptation (2020, neutral): “1 TPN [35] 93.”
MIC: Masked Image Consistency for Context-Enhanced Domain Adaptation (2022, methodology): “Due to the ubiquity of domain gaps, UDA methods were designed for all major computer vision problems including image classification [20, 55, 56, 64], semantic segmentation [30, 32, 84, 102], and object detection [7, 8, 49, 52].”