Unsupervised Visual Domain Adaptation Using Subspace Alignment

By Basura Fernando, Amaury Habrard, Marc Sebban, Tinne Tuytelaars. 2013

1,373 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Cancer Research, 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 Cancer Research

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

Citations in the corpus

Unsupervised Image-to-Image Translation Networks (2017, background): “Early UDA works have explored ideas from subspace learning (Fernando et al., 2013) to deep learning (Ganin et al.”
Domain-Adversarial Training of Neural Networks (2015, methodology): “At the same time, the improvement over source-only model achieved by subspace alignment (SA) (Fernando et al., 2013) is quite modest, thus highlighting the diﬃculty of the adaptation task.”
Few-Shot Adversarial Domain Adaptation (2017, background): “Some try to find a function which can map from the source domain to the target domain [47, 28, 19, 16, 11, 55, 52].”
Transductive Multi-View Zero-Shot Learning (2015, background): “This process alleviates the projection domain shift problem, as well as providing a common space in which heterogeneous views can be directly compared, and their complementarity exploited (Sec.”
Simultaneous Deep Transfer Across Domains and Tasks (2015, background): “A large number of methods have sought to overcome this difference by learning a feature space transformation to align the source and target representations [28, 23, 11, 15].”
Deep Domain Confusion: Maximizing for Domain Invariance (2014, neutral): “learned representations. Our method then selects the dimensionality that minimizes the MMD between the source and target data. A!W D!W W!D Average GFK(PLS,PCA) [16] 46.4 0.5 61.3 0.4 66.3 0.4 53.0 SA [13] 45.0 64.8 69.9 59.9 DA-NBNN [31] 52.8 3.7 76.6 1.7 76.2 2.5 68.5 DLID [8] 51.9 78.2 89.9 73.3 D”
Taskonomy: Disentangling Task Transfer Learning (2018, background): “Domain adaption seeks to render a function that is developed on a certain domain applicable to another [42, 96, 5, 77, 50, 24, 34].”
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, methodology): “The features are fixed and then a model is trained to align the source and target domains [16, 20, 33, 38, 42, 43, 44].”
Deep CORAL: Correlation Alignment for Deep Domain Adaptation (2016, methodology): “We compare to 7 recently published methods: CNN [20] (no adaptation), GFK [7], SA [8], TCA [22], CORAL [1], DDC [16], DAN [17].”
One-Shot Adaptation of Supervised Deep Convolutional Models (2013, methodology): “In this scenario, we apply two state-of-the-art unsupervised adaptation methods, GFK [12] and SA [10].”
Unified Deep Supervised Domain Adaptation and Generalization (2017, background): “The first one includes those that try to find a mapping between source and target distributions [53, 34, 27, 26, 19, 57].”
Domain Adaptation for Visual Applications: A Comprehensive Survey (2017, methodology): “ in interpolating between domains based on the amount of source and target data used to train each model. (Image courtesy S. Chopra [128]). were applied to DECAF features in [123], Subspace Alignment [19] and Correlation Alignment in [21]. [70] experiments with DeCAF features within the extended MDA”
Return of Frustratingly Easy Domain Adaptation (2015, methodology): “As demonstrated in (Kulis, Saenko, and Darrell 2011; Harel and Mannor 2011; Fernando et al. 2013), asymmetric transformations are more flexible and often yield better performance for domain adaptation tasks.”
A Deeper Look at Dataset Bias (2015, methodology): “This approach learns an alignment matrix A between the source and the target subspace by minimizing the following Bregman divergence [12]: Here X s and X t are the subspace bases obtained by applying PCA over the source and target data and selecting for each domain the d eigenvectors corresponding…”