Domain Adaptation on the Statistical Manifold

Baktashmotlagh, Mahsa; Harandi, Mehrtash; Lovell, Brian C.; Salzmann, Mathieu

doi:10.1109/cvpr.2014.318

Cited by 129 publications

(69 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similarly, Baktashmotlagh et al [18] propose a statistically invariant sample selection method to choose landmarks using the Hellinger distance instead of MMD. However, in their approaches, the landmarks are selected from the source domain, and are not necessarily the samples close to the target domain, while our approach identifies the target instances, which are more likely to be correctly predicted by the source classifier.…”

Section: Related Workmentioning

confidence: 98%

See 1 more Smart Citation

Prediction Reweighting for Domain Adaptation

Song

Huang

2017

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

There are plenty of classification methods that perform well when training and testing data are drawn from the same distribution. However, in real applications, this condition may be violated, which causes degradation of classification accuracy. Domain adaptation is an effective approach to address this problem. In this paper, we propose a general domain adaptation framework from the perspective of prediction reweighting, from which a novel approach is derived. Different from the major domain adaptation methods, our idea is to reweight predictions of the training classifier on testing data according to their signed distance to the domain separator, which is a classifier that distinguishes training data (from source domain) and testing data (from target domain). We then propagate the labels of target instances with larger weights to ones with smaller weights by introducing a manifold regularization method. It can be proved that our reweighting scheme effectively brings the source and target domains closer to each other in an appropriate sense, such that classification in target domain becomes easier. The proposed method can be implemented efficiently by a simple two-stage algorithm, and the target classifier has a closed-form solution. The effectiveness of our approach is verified by the experiments on artificial datasets and two standard benchmarks, a visual object recognition task and a cross-domain sentiment analysis of text. Experimental results demonstrate that our method is competitive with the state-of-the-art domain adaptation algorithms.

show abstract

Section: Related Workmentioning

confidence: 98%

“…By utilizing (18), it is necessary to have the knowledge of the distribution of Z to calculate ρ. According to the definition of Z , it is a linear projection from high dimensional to 1-D space, essentially.…”

Section: Justificationmentioning

confidence: 99%

Prediction Reweighting for Domain Adaptation

Song

Huang

2017

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

show abstract

“…Geodesic flow kernel (GFK) [25] extended the idea of sampled points in manifold [27] and proposed to learn the geodesic flow kernel between domains. The work of [3] used a Hellinger distance to approximate the geodesic distance in Riemann space. [2] proposed to use Grassmann for domain adaptation, but they ignored the conditional distribution alignment.…”

Section: Subspace and Manifold Learningmentioning

confidence: 99%

“…However, in this paper, we argue that this assumption is not practical in real applications. For example, when two domains are very dissimilar (e.g., transfer learning between (1) and (3) in Fig. 1(a)), the marginal distribution is more important.…”

Section: Introductionmentioning

confidence: 99%

Transfer Learning with Dynamic Distribution Adaptation

Wang

Chen

Feng

et al. 2020

ACM Trans. Intell. Syst. Technol.

172

104

View full text Add to dashboard Cite

Transfer learning aims to learn robust classifiers for the target domain by leveraging knowledge from a source domain. Since the source and the target domains are usually from different distributions, existing methods mainly focus on adapting the cross-domain marginal or conditional distributions. However, in real applications, the marginal and conditional distributions usually have different contributions to the domain discrepancy. Existing methods fail to quantitatively evaluate the different importance of these two distributions, which will result in unsatisfactory transfer performance. In this paper, we propose a novel concept called Dynamic Distribution Adaptation (DDA), which is capable of quantitatively evaluating the relative importance of each distribution. DDA can be easily incorporated into the framework of structural risk minimization to solve transfer learning problems. On the basis of DDA, we propose two novel learning algorithms: (1) Manifold Dynamic Distribution Adaptation (MDDA) for traditional transfer learning, and (2) Dynamic Distribution Adaptation Network (DDAN) for deep transfer learning. Extensive experiments demonstrate that MDDA and DDAN significantly improve the transfer learning performance and setup a strong baseline over the latest deep and adversarial methods on digits recognition, sentiment analysis, and image classification. More importantly, it is shown that marginal and conditional distributions have different contributions to the domain divergence, and our DDA is able to provide good quantitative evaluation of their relative importance which leads to better performance. We believe this observation can be helpful for future research in transfer learning.

show abstract

“…Semisupervised domain adaptation usually constructs learning model using the priori information from the class labels of the source and target samples [11], or the pairwise similarity between them [3,4]. As no class label nor similarity information of the source and target samples is available, unsupervised domain adaptation often exploits the underlying geometry structure [15,16] or the intrinsic probability distribution of data across domains to bridge the gap of domains [9,17]. In this paper, we focus on the issue of unsupervised visual domain adaptation, which is more common in real world applications and relatively more challenging as well.…”

Section: Introductionmentioning

confidence: 99%