Contrastive Adaptation Network for Single- and Multi-Source Domain Adaptation

Abstract

Unsupervised Domain Adaptation (UDA) makes predictions for the target domain data while manual annotations are only available in the source domain. Previous methods minimize the domain discrepancy neglecting the class information, which may lead to misalignment and poor generalization performance. To tackle this issue, this paper proposes Contrastive Adaptation Network (CAN) that optimizes a new metric named Contrastive Domain Discrepancy explicitly modeling the intra-class domain discrepancy and the inter-class domain discrepancy. To optimize CAN, two technical issues need to be addressed: 1) the target labels are not available and 2) the conventional mini-batch sampling is imbalanced. Thus we design an alternating update strategy to optimize both the target label estimations and the feature representations. Moreover, we develop class-aware sampling to enable more efficient and effective training. Our framework can be generally applied to the single-source and multi-source domain adaptation scenarios. In particular, to deal with multiple source domain data, we propose 1) multi-source clustering ensemble which exploits the complementary knowledge of distinct source domains to make more accurate and robust target label estimations, and 2) boundary-sensitive alignment to make the decision boundary better fitted to the target. Experiments are conducted on three real-world benchmarks (i.e., Office-31 and VisDA-2017 for the single-source scenario, DomainNet for the multi-source scenario). All the results demonstrate that our CAN performs favorably against the state-of-the-art methods. Ablation studies also verify the effectiveness of each key component of our proposed system.

Cross-Site Severity Assessment of COVID-19 From CT Images via Domain Adaptation

“…Domain shifts may exist not only between D s,e and D t , but also between the source domains. Hence, we assign each source domain D s,e a different domain translator D e (•) and a classifier L e (•), whereas the feature extractor F (•) is shared across all domains to learn domain-invariant features as [51]- [53]. For the target domain, data X u is forwarded into F (•) to extract feature Z u , which is then fed into the multiple classifiers {L e (•)}.…”

“…Table VIII shows the comparisons of M-FLARE with some recent methods on the multi-source DA task (S1,S2,S3)→S1. Following [51]- [53], we introduce two standards: 1) singlebest domain adaptation, which reports the single-source adaptation result best-performing in the test set, and 2) source combine domain adaptation, which reports the single-source adaptation result by transferring the combination of multiple source domains to a target domain. The first standard evaluates whether the single-source adaptation performance can be improved by introducing other source domains; the second testifies the need of exploiting a multi-source DA…”

Overcoming the Domain Gap in Contrastive Learning of Neural Action Representations

“…Maximum Mean Discrepancy (MMD): We replaced adversarial loss with a statistical test to minimize the distributional discrepancy from different domains [67]. Similar to previous work, we applied MMD only on the representations before the projection layer independently on both modalities [35,50]. Similar to the GLR baseline, we first trained 10 epochs only using the contrastive loss, and trained using the combined losses L N CE + λ M M D L M M D for the remainder.…”

“…However, labeling behavioral-neural datasets requires expensive and arduous manual labor by trained scientists, and thus often leaving the vast majority of data unlabeled. Similarly, it is non-trivial to generalize few-shot domain adaptation methods to multimodal tasks [50,51]. Thus, the field of neuroscience needs new computational approaches that can extract information from ever-increasing amounts of unlabeled multimodal datasets that also suffer from extensive domain gaps across subjects.…”

Domain Attention Consistency for Multi-Source Domain Adaptation

“…This problem has persisted in the deep learning era, even when deep convolutional neural networks (CNNs) have demonstrated great successes in solving many recognition tasks [32,50]. As a key solution to overcome the domain shift problem, unsupervised domain adaptation (UDA) has been extensively studied [2,8,9,16,23,33,49]. UDA aims to transfer the knowledge learned from one or multiple labeled source domains to a target domain in which only unlabeled data are given for model adaptation.…”