Learning with Augmented Features for Heterogeneous Domain Adaptation

By Lixin Duan, Dong Xu, Ivor Wai-Hung Tsang. 2012

357 citations recorded by the source corpus.

We propose a new learning method for heterogeneous domain adaptation (HDA), in which the data from the source domain and the target domain are represented by heterogeneous features with different dimensions. Using two different projection matrices, we first transform the data from two domains into a common subspace in order to measure the similarity between the data from two domains. We then propose two new feature mapping functions to augment the transformed data with their original features and zeros. The existing learning methods (e.g., SVM and SVR) can be readily incorporated with our newly proposed augmented feature representations to effectively utilize the data from both domains for HDA. Using the hinge loss function in SVM as an example, we introduce the detailed objective function in our method called Heterogeneous Feature Augmentation (HFA) for a linear case and also describe its kernelization in order to efficiently cope with the data with very high dimensions. Moreover, we also develop an alternating optimization algorithm to effectively solve the nontrivial optimization problem in our HFA method. Comprehensive experiments on two benchmark datasets clearly demonstrate that HFA outperforms the existing HDA methods.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

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 Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

Citations in the corpus

Domain randomization for transferring deep neural networks from simulation to the real world (2017, background): “The computer vision community has devoted signiﬁcant study to the problem of adapting vision-based models trained in a source domain to a previously unseen target domain (see, e.g., [9], [14], [15], [19], [23], [25], [51]).”
One-Shot Imitation Learning (2017, background): “Success stories include domain adaptation in computer vision [64, 34, 28, 4, 15, 24, 33, 59, 14] and control [60, 45, 46, 20, 54].”
Multi-domain Neural Network Language Generation for Spoken Dialogue Systems (2016, methodology): “Recently, RNN-based language generation has been introduced (Wen et al., 2015a; Wen et al., 2015b).”
Cross Modal Distillation for Supervision Transfer (2015, methodology): “In this section we present experimental results for the NYUD2 dataset where we use color and depth images as the paired modalities, and on the JHMDB video dataset for which we use the RGB and optical flow frames as the two modalities.”
Simultaneous Deep Transfer Across Domains and Tasks (2015, background): “Others have sought to both learn a feature transformation and regularize a target classiﬁer simultaneously [18, 10].”
Deep Domain Confusion: Maximizing for Domain Invariance (2014, background): “Others have sought to both learn a feature transformation and regularize a target classiﬁer simultaneously [20, 12].”
Third-Person Imitation Learning (2017, background): “…of adapting what is learned in one domain to another domain has been studied extensively in computer vision in the supervised learning setting Yang et al. (2007); Mansour et al. (2009); Kulis et al. (2011); Aytar & Zisserman (2011); Duan et al. (2012); Hoffman et al. (2013); Long & Wang (2015).”
LSDA: Large Scale Detection through Adaptation (2014, background): “, feature space transformations [12], model adaptation approaches [13, 14] and joint feature and model adaptation [15, 16].”
A Comprehensive Survey on Transfer Learning (2019, background): “To solve this problem, Li et al. [67] proposed an approach termed heterogeneous feature augmentation (HFA) [66].”
Domain Adaptation for Visual Applications: A Comprehensive Survey (2017, methodology): “re multi-source HDA approaches. 23 Code available at http://www.umiacs.umd.edu/˜pvishalm/Codes/DomainAdaptDict.zip Domain Adaptation for Visual Applications: A Comprehensive Survey 19 Fig. 15 The HFA [114] is seeking for an optimal common space while simultaneously learning a discriminative SVM clas”
Adapting Deep Visuomotor Representations with Weak Pairwise Constraints (2015, background): “t al., 2007; Aytar & Zisserman, 2011). Further models improved upon this by proposing adaptation which both transformed the representation spaces and regularized the target model using the source Duan et al. (2012); Hoffman et al. (2013). Since the resurgence in the popularity of convolutional n”
Transfer learning for speech and language processing (2015, methodology): “factorization [17], RBM-based latent factor learning [27] , or joint transfer optimization [28], [16], [29].”
Semi-supervised Domain Adaptation with Instance Constraints (2013, methodology): “Supervised approaches do this by utilizing a few labeled examples in the target domain [16, 1, 7, 9, 18, 24, 28], while semi-supervised methods also take into account the (typically much more abundant) unlabeled target samples [13, 14].”
Towards Adapting Deep Visuomotor Representations from Simulated to Real Environments (2015, methodology): “Further models improved upon this by proposing adaptation which both transformed the representation spaces and regularized the target model using the source [23], [24].”
Fast Generalized Distillation for Semi-Supervised Domain Adaptation (2017, methodology): “Previous work has addressed the importance of knowledge control in domain adaptation (Duan, Xu, and Tsang 2012; Duan et al. 2012).”
Recognizing RGB Images by Learning from RGB-D Data (2014, methodology): “However, labeled samples in the target domain must be provided in the existing HDA methods [9, 24, 22], while we do not require any labeled target domain samples in this work.”
Heterogeneous Domain Adaptation for Multiple Classes (2014, methodology): “[15] proposed a Heterogeneous Feature Augmentation (HFA) method to augment homogeneous common features that learned by a maximum-margin approach from both the source and target domains.”