Connecting the Dots with Landmarks: Discriminatively Learning Domain-Invariant Features for Unsupervised Domain Adaptation

By Boqing Gong, Kristen Grauman, Fei Sha. 2013

521 citations recorded by the source corpus.

Subfields: 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 Representation & Self-Supervised Learning

Citations in the corpus

Domain-Adversarial Training of Neural Networks (2015, methodology): “Our training protocol is adopted from Gong et al. (2013); Chopra et al. (2013); Long and Wang (2015) as during adaptation we use all available labeled source examples and unlabeled target examples (the premise of our method is the abundance of unlabeled data in the target domain).”
Multiple-Source Adaptation with Domain Classifiers (2020, methodology): “…target domain [Mansour et al., 2008, 2009a, Hoffman et al., 2018, Pan and Yang, 2010, Muandet et al., 2013, Xu et al., 2014, Hoffman et al., 2012, Gong et al., 2013a,b, Zhang et al., 2015, Ganin et al., 2016, Tzeng et al., 2015, Motiian et al., 2017b,a, Wang et al., 2019, Konstantinov and…”
A Theory of Multiple-Source Adaptation with Limited Target Labeled Data (2020, methodology): “This approach has been further used in many applications such as object recognition (Ho↵man et al., 2012; Gong et al., 2013a,b).”
Agnostic Federated Learning (2019, background): “…more closely related to that of federated learning is that of multiple-source adaptation , ﬁrst formalized and analyzed theoretically by Mansour, Mohri, and Rostamizadeh (2009c,a) and later studied for various applications such as object recognition (Hoffman et al., 2012; Gong et al., 2013a,b).”
Algorithms and Theory for Multiple-Source Adaptation (2018, background): “This is known as the multiple-source adaption (MSA) problem ﬁrst formalized and analyzed theoretically by Mansour et al. [2008, 2009] and later studied for various applications such as object recognition [Hoffman et al., 2012, Gong et al., 2013a,b].”
Domain Adaptation with Conditional Transferable Components (2016, methodology): “In our experiments, we use the evaluation protocol in (Gong et al., 2013).”
FastMMD: Ensemble of Circular Discrepancy for Efficient Two-Sample Test (2014, background): “…2010; 2011; Sejdinovic et al., 2013) and successful applications including biological data test (Borgwardt et al., 2006), data integration and attribute matching (Gretton et al., 2009), outlier detection, data classifiability (Sriperumbudur et al., 2009), domain adaption (Gong et al., 2013), etc.”
One-Shot Imitation from Observing Humans via Domain-Adaptive Meta-Learning (2018, background): “Many of these techniques aim to find a representation that is invariant to the domain [12, 16, 54, 41, 25].”
Deep Hashing Network for Unsupervised Domain Adaptation (2017, background): “Some other procedures re-weight the data points in the source domain, to select source data that is similar to the target, when training a domain adaptive classifier, [9, 10, 19].”
Domain Adaptation for Visual Applications: A Comprehensive Survey (2017, methodology): “available labeled set and using the new cross-domain representation, either a discriminative classiﬁer is trained [107] or the target labels are deduced by nearest neighbor search in the kernel space [103, 101]. Feature space alignment. Instead of of augmenting the features, other methods tries to a”
Return of Frustratingly Easy Domain Adaptation (2015, methodology): “We use the processed data from (Gong, Grauman, and Sha 2013), in which the dimensionality of the bag-of-words features was reduced to keep the top 400 words without losing performance.”
A Deeper Look at Dataset Bias (2015, methodology): “The landmark method introduced in [13] allows to sample the source training images that are more relevant for the target test data.”
Deep Reconstruction-Classification Networks for Unsupervised Domain Adaptation (2016, methodology): “We followed the standard unsupervised domain adaptation training protocol used in previous work [38,7,39], that is, using all labeled source data and unlabeled target data.”
Revisiting Batch Normalization For Practical Domain Adaptation (2016, methodology): “All of these methods face the same challenge of devising an eﬀective domain transfer function in high-dimensional non-linear space.”
Transferability vs. Discriminability: Batch Spectral Penalization for Adversarial Domain Adaptation (2019, background): “Early domain adaptation methods in the shallow regime learn feature representations invariant across domains (Pan et al., 2011; Gong et al., 2012) or reweigh source instances based on their relevance to the target domain (Huang et al., 2007; Gong et al., 2013).”