Domain Adaptation with Regularized Optimal Transport

By Nicolas Courty, Rémi Flamary, Devis Tuia. 2014

249 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computational Mechanics, 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

1. Mohammad Rostami
2. Mingsheng Long
3. Jianmin Wang
4. Dacheng Tao
5. Jiashi Feng
6. Kate Saenko
7. Yi Yang
8. Trevor Darrell

In Computational Mechanics

1. Marco Cuturi
2. Alain Rakotomamonjy
3. G. Peyr'e
4. Francis R. Bach
5. Soheil Kolouri
6. Lénaïc Chizat
7. Aude Genevay
8. Gustavo Kunde Rohde

In Computer Vision and Pattern Recognition

1. Mohammad Rostami
2. Luc Van Gool
3. Dacheng Tao
4. Jiashi Feng
5. Trevor Darrell
6. Xiaogang Wang
7. Yi Yang
8. Ming-Hsuan Yang

Citations in the corpus

• Learning Generative Models with Sinkhorn Divergences (2017, methodology): “The benefits of this regularization has opened the path to many applications of the Wasserstein distance in relevant learning problems [8, 13, 18, 28].”
• Optimal Transport for Domain Adaptation (2014, positive): “We also present some results with our previous class-label based regularizer built upon an (cid:96) p − (cid:96) 1 norm: OT-LpL1 [14].”
• Optimal Mass Transport: Signal processing and machine-learning applications (2017, methodology): “Some examples covered in this article include image retrieval [32], [44], signal and image representation [25], [27], [40], [50], inverse problems [30], cancer detection [4], [39], texture and color modeling [18], [41], shape and image registration [22], [29], and machine learning [12], [17], [19], ”
• Reliable Weighted Optimal Transport for Unsupervised Domain Adaptation (2020, methodology): “A sophisticated approach is to formally define a statistical distance in the prob-4395 abilistic metric space and learn the optimal transport coupling to minimize that distance [7, 5, 33].”
• Gromov-Wasserstein Averaging of Kernel and Distance Matrices (2016, background): “, 2014), domain adaptation (Courty et al., 2014), designing loss functions (Frogner et al.”
• Joint distribution optimal transportation for domain adaptation (2017, methodology): “Optimal transport has been used in DA as a principled way to bring the source and target distribution closer [13, 14, 17], by seeking for a transport plan between the empirical distributions of Xs and Xt and interpolating Xs thanks to a barycentric mapping [14], or by estimating a mapping which is n”
• Wasserstein Distance Guided Representation Learning for Domain Adaptation (2017, background): “Another related work for domain adaptation is optimal transport (Courty, Flamary, and Tuia 2014; Courty et al. 2017), which is equivalent to Wasserstein distance.”
• Sample Complexity of Sinkhorn Divergences (2018, background): “Among the many learning tasks performed with this regularization, one may cite domain adaptation (Courty et al., 2014), text retrieval (Kus-ner et al., 2015) or multi-label classification (Frogner et al., 2015).”
• DeepJDOT: Deep Joint distribution optimal transport for unsupervised domain adaptation (2018, unknown): “Optimal transport [22–24] has been used in domain adaptation to learn the transformation between domains [4, 25, 26], with associated theoretical guarantees [27].”