What you saw is not what you get: Domain adaptation using asymmetric kernel transforms

By Brian Kulis, Kate Saenko, Trevor Darrell. 2011

767 citations recorded by the source corpus.

In real-world applications, “what you saw” during training is often not “what you get” during deployment: the distribution and even the type and dimensionality of features can change from one dataset to the next. In this paper, we address the problem of visual domain adaptation for transferring object models from one dataset or visual domain to another. We introduce ARC-t, a flexible model for supervised learning of non-linear transformations between domains. Our method is based on a novel theoretical result demonstrating that such transformations can be learned in kernel space. Unlike existing work, our model is not restricted to symmetric transformations, nor to features of the same type and dimensionality, making it applicable to a significantly wider set of adaptation scenarios than previous methods. Furthermore, the method can be applied to categories that were not available during training. We demonstrate the ability of our method to adapt object recognition models under a variety of situations, such as differing imaging conditions, feature types and codebooks.

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]).”
Predicting Deep Zero-Shot Convolutional Neural Networks Using Textual Descriptions (2015, methodology): “(3) is equivalent to the low rank matrix factorization [15].”
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].”
DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition (2013, methodology): “We evaluate on the SUN-397 scene dataset, as well as datasets used to evaluate domain adaptation performance directly (Chopra et al., 2013; Kulis et al., 2011).”
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].”
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): “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, 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 [29, 25, 13, 16].”
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].”
Undoing the Damage of Dataset Bias (2012, methodology): “[8] generalize the method in [7] to handle asymmetric transformations in which the feature dimensionality in source and target domain can be different.”
Efficient Learning of Domain-invariant Image Representations (2013, methodology): “For visual domain adaptation, transform-based adaptation methods [1, 12, 13, 2, 14, 11] have recently been proposed.”
One-Shot Adaptation of Supervised Deep Convolutional Models (2013, methodology): “Many visual domain adaptation methods have been put forth to compensate for dataset bias [7, 22, 1, 20, 18, 16, 13, 12, 14, 15], but are limited to shallow models.”
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): “Hoffmann et al . propose in [116] a constrained clustering algorithm 17 to identify automatically ”source” domains in a large data set, on which they apply a multi-source extension of the Asymmetric Kernel Transforms (AST) [129].”
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.”
Learning with Augmented Features for Heterogeneous Domain Adaptation (2012, methodology): “The proof can be analogously derived as for Lemma 3.1 in (Kulis et al., 2011).”
Pixel-Level Domain Transfer (2016, methodology): “different image domain has been proposed in computer vision [9,10,11,12,13,14], but all these adaptations take place in the feature space, i.”