mlsanityData to 13 Oct 2026

Unbalanced minibatch Optimal Transport; applications to Domain Adaptation

arXiv 2103.03606 ↗First posted 5 Mar 2021cs.LG math.ST stat.MLcited by 123

Not on the map. Why?

Abstract (arXiv)

Optimal transport distances have found many applications in machine learning for their capacity to compare non-parametric probability distributions. Yet their algorithmic complexity generally prevents their direct use on large scale datasets. Among the possible strategies to alleviate this issue, practitioners can rely on computing estimates of these distances over subsets of data, em i.e. minibatches. While computationally appealing, we highlight in this paper some limits of this strategy, arguing it can lead to undesirable smoothing effects. As an alternative, we suggest that the same minibatch strategy coupled with unbalanced optimal transport can yield more robust behavior. We discuss the associated theoretical properties, such as unbiased estimators, existence of gradients and concentration bounds. Our experimental study shows that in challenging problems associated to domain adaptation, the use of unbalanced optimal transport leads to significantly better results, competing with or surpassing recent baselines.

Key facts. Cited by 123 papers in the map (data through 13 Oct 2026). Builds on 35 of its 64 references that are in the map. First posted on arXiv 5 Mar 2021.

Builds on

35 of 64 references.

References · 35

  1. 2015Domain-Adversarial Training of Neural Networkscited by 3,030
  2. 2013Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distancescited by 2,295
  3. 2017Improved Training of Wasserstein GANscited by 2,226
  4. 2017Deep Hashing Network for Unsupervised Domain Adaptationcited by 1,183
  5. 2019When Does Label Smoothing Help?cited by 986
  6. 2018Computational Optimal Transportcited by 772
  7. 2015Optimal Transport for Domain Adaptationcited by 460
  8. 2017VisDA: The Visual Domain Adaptation Challengecited by 434
  9. 2017Learning Generative Models with Sinkhorn Divergencescited by 311
  10. 2015Data-driven Distributionally Robust Optimization Using the Wasserstein Metric: Performance Guarantees and Tractable Reformulationscited by 287
  11. 2017Joint Distribution Optimal Transportation for Domain Adaptationcited by 281
  12. 2018DeepJDOT: Deep Joint Distribution Optimal Transport for Unsupervised Domain Adaptationcited by 213
23 more
  1. 2019Wasserstein Distributionally Robust Optimization: Theory and Applications in Machine Learningcited by 170
  2. 2015Optimal Entropy-Transport problems and a new Hellinger-Kantorovich distance between positive measurescited by 138
  3. 2018Partial Adversarial Domain Adaptationcited by 127
  4. 2018Stochastic subgradient method converges on tame functionscited by 121
  5. 2017The Cramer Distance as a Solution to Biased Wasserstein Gradientscited by 116
  6. 2020Missing Data Imputation using Optimal Transportcited by 99
  7. 2019Subspace Robust Wasserstein Distancescited by 90
  8. 2019Learning to Transfer Examples for Partial Domain Adaptationcited by 86
  9. 2020Robust Optimal Transport with Applications in Generative Modeling and Domain Adaptationcited by 86
  10. 2019Learning with minibatch Wasserstein : asymptotic and gradient propertiescited by 84
  11. 2018Sliced-Wasserstein Flows: Nonparametric Generative Modeling via Optimal Transport and Diffusionscited by 77
  12. 2020Adversarial-Learned Loss for Domain Adaptationcited by 65
  13. 2015Sliced Wasserstein Kernels for Probability Distributionscited by 64
  14. 2021Minibatch optimal transport distances; analysis and applicationscited by 64
  15. 2020On Unbalanced Optimal Transport: An Analysis of Sinkhorn Algorithmcited by 61
  16. 2020A Balanced and Uncertainty-aware Approach for Partial Domain Adaptationcited by 55
  17. 2019Sinkhorn Divergences for Unbalanced Optimal Transportcited by 52
  18. 2020Outlier-Robust Optimal Transportcited by 45
  19. 2018Optimal Transport: Fast Probabilistic Approximation with Exact Solverscited by 27
  20. 2018Analysis of nonsmooth stochastic approximation: the differential inclusion approachcited by 23
  21. 2013On properties of the Generalized Wasserstein distancecited by 11
  22. 2018An Entropic Optimal Transport Loss for Learning Deep Neural Networks under Label Noise in Remote Sensing Imagescited by 11
  23. 2017A Framework for Wasserstein-1-Type Metricscited by 8

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2021: 11; 2022: 24; 2023: 31; 2024: 19; 2025: 21; 2026: 17.

123 papers.

Sort:

By area:

  1. 2023Improving and generalizing flow-based generative models with minibatch optimal transportcited by 857
  2. 2023SE(3)-Stochastic Flow Matching for Protein Backbone Generationcited by 148
  3. 2023Multimodal Optimal Transport-based Co-Attention Transformer with Global Structure Consistency for Survival Predictioncited by 86
  4. 2023Simulation-free Schr"odinger bridges via score and flow matchingcited by 81
  5. 2021LCDNet: Deep Loop Closure Detection and Point Cloud Registration for LiDAR SLAMcited by 67
  6. 2023Recent Advances in Optimal Transport for Machine Learningcited by 66
  7. 2022Unified Optimal Transport Framework for Universal Domain Adaptationcited by 57
  8. 2021DINE: Domain Adaptation from Single and Multiple Black-box Predictorscited by 51
  9. 2022Unbalanced Optimal Transport, from Theory to Numericscited by 45
  10. 2024Global and Local Prompts Cooperation via Optimal Transport for Federated Learningcited by 45
  11. 2021Improving Mini-batch Optimal Transport via Partial Transportationcited by 40
  12. 2023Generative Modeling through the Semi-dual Formulation of Unbalanced Optimal Transportcited by 40
38 more
  1. 2023Optimal Transport for Treatment Effect Estimationcited by 32
  2. 2021Outlier-Robust Optimal Transport: Duality, Structure, and Statistical Analysiscited by 27
  3. 2023Scalable Optimal Transport Methods in Machine Learning: A Contemporary Surveycited by 27
  4. 2024Addressing Misspecification in Simulation-based Inference through Data-driven Calibrationcited by 27
  5. 2021UMAD: Universal Model Adaptation under Domain and Category Shiftcited by 23
  6. 2022Faster Unbalanced Optimal Transport: Translation invariant Sinkhorn and 1-D Frank-Wolfecited by 22
  7. 2021TimeMatch: Unsupervised Cross-Region Adaptation by Temporal Shift Estimationcited by 21
  8. 2023SALUDA: Surface-based Automotive Lidar Unsupervised Domain Adaptationcited by 21
  9. 2022Multi-level Consistency Learning for Semi-supervised Domain Adaptationcited by 20
  10. 2022Sliced Optimal Partial Transportcited by 20
  11. 2021Unsupervised Domain Adaptation for LiDAR Panoptic Segmentationcited by 19
  12. 2022Divide to Adapt: Mitigating Confirmation Bias for Domain Adaptation of Black-Box Predictorscited by 19
  13. 2023GenPhys: From Physical Processes to Generative Modelscited by 19
  14. 2022Sobolev Transport: A Scalable Metric for Probability Measures with Graph Metricscited by 18
  15. 2022Neural Unbalanced Optimal Transport via Cycle-Consistent Semi-Couplingscited by 17
  16. 2022On Unbalanced Optimal Transport: Gradient Methods, Sparsity and Approximation Errorcited by 16
  17. 2021k-Mixup Regularization for Deep Learning via Optimal Transportcited by 14
  18. 2023Unbalanced Optimal Transport: A Unified Framework for Object Detectioncited by 14
  19. 2022Continual Variational Autoencoder Learning via Online Cooperative Memorizationcited by 12
  20. 2024Accelerating Sinkhorn Algorithm with Sparse Newton Iterationscited by 11
  21. 2023Scalable Unbalanced Sobolev Transport for Measures on a Graphcited by 10
  22. 2022Improving Molecular Representation Learning with Metric Learning-enhanced Optimal Transportcited by 9
  23. 2023Semi-supervised Domain Adaptation via Prototype-based Multi-level Learningcited by 9
  24. 2024Generalized Sobolev Transport for Probability Measures on a Graphcited by 9
  25. 2025Hierarchical Refinement: Optimal Transport to Infinity and Beyondcited by 9
  26. 2025Generalizable Domain Adaptation for Sim-and-Real Policy Co-Trainingcited by 9
  27. 2022Distributional Reinforcement Learning with Regularized Wasserstein Losscited by 8
  28. 2023Robust Estimation under the Wasserstein Distancecited by 8
  29. 2023Linear Optimal Partial Transport Embeddingcited by 8
  30. 2023Light Unbalanced Optimal Transportcited by 8
  31. 2023Analyzing and Improving Optimal-Transport-based Adversarial Networkscited by 8
  32. 2024Neural Sinkhorn Gradient Flowcited by 8
  33. 2023Unbalanced Low-rank Optimal Transport Solverscited by 7
  34. 2023Optimal Transport for Measures with Noisy Tree Metriccited by 7
  35. 2024Prototypical Partial Optimal Transport for Universal Domain Adaptationcited by 7
  36. 2022Unbalanced CO-Optimal Transportcited by 6
  37. 2023Consistent Optimal Transport with Empirical Conditional Measurescited by 6
  38. 2024Domain Adaptation with Cauchy-Schwarz Divergencecited by 6

50 most cited of 123 All 123 (JSON)