Cross-domain video concept detection using adaptive svms

By Jun Yang, Rong Yan, Alexander G. Hauptmann. 2007

747 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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

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].”
Agnostic Federated Learning (2019, background): “…unsupervised adaptation techniques (Gong et al., 2012; Long et al., 2015; Ganin and Lempitsky, 2015; Tzeng et al., 2015), or via lightly supervised ones (some amount of labeled data from the target domain) (Saenko et al., 2010; Yang et al., 2007; Hoffman et al., 2013; Girshick et al., 2014).”
Algorithms and Theory for Multiple-Source Adaptation (2018, background): “…et al., 2012, Long et al., 2015, Ganin and Lempitsky, 2015, Tzeng et al., 2015], or via lightly supervised ones (some amount of labeled data from the target domain) [Saenko et al., 2010, Yang et al., 2007, Hoffman et al., 2013, Girshick et al., 2014], has been extensively investigated in the past.”
A PAC-Bayesian bound for Lifelong Learning (2013, methodology): “It can be captured by regularizing the distance to this vector (Aytar & Zisserman, 2011; Yang et al., 2007): where w pr is some function of weight vectors of previously observed tasks, e.g. just their average, w pr = 1 n =1 w i .”
Third-Person Imitation Learning (2017, background): “…ﬁnd that this simple approach has been able to solve the problems that are presented in this paper (illustrated in Figure 1), even though the student’s observations are related in a complicated way to the teacher’s demonstrations (given that the observations and the demonstrations are pixel-level).”
LSDA: Large Scale Detection through Adaptation (2014, methodology): “These include variants for linear support vector machines [17, 18, 19], as well as adaptive latent SVMs [20] and adaptive exemplar SVM [21].”
Curriculum learning of multiple tasks (2014, methodology): “For a given weight vector w̃ and training data for a task, the Adaptive SVM performs the following optimization: min w ‖w − w̃‖2 + C m m∑ j=1 ξj (2) sb.t. yj〈w, xj〉 ≥ 1− ξj , ξj ≥ 0 for all 1 ≤ j ≤ m. Specifically, to learn a predictor for the task tπ(i) we solve (2) using the weight vector obtained”
Domain Adaptation for Visual Applications: A Comprehensive Survey (2017, methodology): “nsductive SVM [123] that aims at decreasing the generalization error of the classiﬁcation, by incorporating knowledge about the target data into the SVM optimization process. The Adaptive SVM (A-SVM) [258] progressively adjusts the decision boundaries of the source classiﬁers with the help of a set ”
Adapting Deep Visuomotor Representations with Weak Pairwise Constraints (2015, methodology): “Traditional visual domain adaptation methods tackled the problem where a fixed representation extraction algorithm was used for both visual domains, and adaptation took the form of learning a transformation between the two spaces [4,15,16] or regularizing the target domain model based on the source ”
SpotTune: Transfer Learning Through Adaptive Fine-Tuning (2018, methodology): “er learning and domain adaptation methods in computer vision [9,37]. Early approaches have concentrated on shallow classiﬁers, using techniques such as instance re-weighting [50,12], model adaptation [11,52], and feature space alignment [44,35,18]. In the multi-task setting, knowing which tasks or p”
Feature-based transfer learning with real-world applications (2010, background): “Applications of transfer learning in these fields include image classification [202, 158, 185, 218, 161, 94, 177], image retrieval [73, 115, 38], face verification from images [196], age estimation from facial images [225], image semantic segmentation [190], video retrieval [208, 73], video concept ”
Write a Classifier: Zero-Shot Learning Using Purely Textual Descriptions (2013, background): “, images or videos from different sources [33, 26, 15, 6].”
Discovering Latent Domains for Multisource Domain Adaptation (2012, methodology): “For example, if the domain specific classifier is (kernelized) nearest neighbors, as in our experiments , and we let x * k (y) be the most similar point in domain k to test point y, then our domain specific object category probability can be expressed as: p(c|d = k, y) = φ A (x * k (y)) T W j φ B (y”
Domain adaptation from multiple sources via auxiliary classifiers (2009, methodology): “However, numerous unlabeled patterns in the target domain are not exploited in the above domain adaptation methods (Wu & Dietterich, 2004; Daumé III, 2007; Yang et al., 2007).”
Semi-supervised Domain Adaptation with Instance Constraints (2013, background): “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].”