Skip to main content
SpringerLink
Log in

Class-rebalanced wasserstein distance for multi-source domain adaptation

  • Published:
Applied Intelligence Aims and scope Submit manuscript

Abstract

In the study of machine learning, multi-source domain adaptation (MSDA) handles multiple datasets which are collected from different distributions by using domain-invariant knowledge extraction. However, the current studies mainly employ features and raw labels on the joint space to perform domain alignment, neglecting the intrinsic structure of label distribution that can harm the performance of adaptation. Therefore, to make better use of label information when aligning joint feature-label distribution, we propose a rebalancing scheme, class-rebalanced Wasserstein distance (CRWD), for unsupervised MSDA under class-wise imbalance and data correlation. Based on the optimal transport for domain adaptation (OTDA) framework, CRWD mitigates the impact of the biased label structure by rectifying the Wasserstein mapping from source to target space. Technically, the class proportions are utilized to encourage distributional transportation between minor classes and principal components, which reweigh the optimal transport plan and reinforce the ground metric of Mahalanobis distance to better metricise the differences among domains. In addition, the scheme measures both inter-domain and intra-source discrepancies to enhance adaptation. Extensive experiments are conducted on various benchmarks, and the results prove that CRWD has competitive advantages.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8

Similar content being viewed by others

Data Availability

The data that support the findings of this study are available from the corresponding author, upon reasonable request.

References

  1. Altschuler J, Niles-Weed J, Rigollet P (2017) Near-linear time approximation algorithms for optimal transport via sinkhorn iteration. Adv Neural Inf Process Syst, vol 30

  2. Arora S, Ge R, Liang Y et al (2017) Generalization and equilibrium in generative adversarial nets (gans). In: International conference on machine learning. PMLR, pp 224–232

  3. Brereton RG (2015) The mahalanobis distance and its relationship to principal component scores. J Chemom 29(3):143–145

    Article  Google Scholar 

  4. Cao Y, Long M, Wang J (2018) Unsupervised domain adaptation with distribution matching machines. In: Proceedings of the AAAI conference on artificial intelligence

  5. Chen Q, Liu Y, Wang Z et al (2018) Re-weighted adversarial adaptation network for unsupervised domain adaptation. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 7976–7985

  6. Courty N, Flamary R, Tuia D, et al. (2016) Optimal transport for domain adaptation. IEEE Trans Pattern Anal Mach Intell 39(9):1853–1865

    Article  Google Scholar 

  7. Courty N, Flamary R, Habrard A et al (2017) Joint distribution optimal transportation for domain adaptation. Adv Neural Inf Process Syst, vol 30

  8. Cuturi M, Avis D (2014) Ground metric learning. J Mach Learn Res 15(1):533–564

    MathSciNet  MATH  Google Scholar 

  9. Damodaran BB, Flamary R, Seguy V, et al. (2020) An entropic optimal transport loss for learning deep neural networks under label noise in remote sensing images. Comp Vision Image Underst 191:102,863

    Article  Google Scholar 

  10. David SB, Lu T, Luu T et al (2010) Impossibility theorems for domain adaptation. In: Proceedings of the thirteenth international conference on artificial intelligence and statistics, JMLR workshop and conference proceedings, pp 129–136

  11. Flamary R, Courty N, Gramfort A, et al. (2021) Pot: python optimal transport. J Mach Learn Res 22(78):1–8

    MATH  Google Scholar 

  12. Gangbo W, Li W, Osher S et al (2019) Unnormalized optimal transport. J Comput Phys 399:108,940

    Article  MathSciNet  MATH  Google Scholar 

  13. Gao P, Wu W, Li J (2021) Multi-source fast transfer learning algorithm based on support vector machine. Appl Intell:1–15

  14. Guo J, Shah D, Barzilay R (2018) Multi-source domain adaptation with mixture of experts. In: Proceedings of the 2018 conference on empirical methods in natural language processing, pp 4694–4703

  15. Guo J, Gong M, Liu T et al (2020) Ltf: A label transformation framework for correcting label shift. In: International conference on machine learning, PMLR, pp 3843–3853

  16. Hu C, Wang Y, Gu J (2020) Cross-domain intelligent fault classification of bearings based on tensor-aligned invariant subspace learning and two-dimensional convolutional neural networks. Knowl-Based Syst 209:106–214

    Article  Google Scholar 

  17. Hu C, He S, Wang Y (2021) A classification method to detect faults in a rotating machinery based on kernelled support tensor machine and multilinear principal component analysis. Appl Intell 51 (4):2609–2621

    Article  Google Scholar 

  18. Li Y, Carlson DE, et al. (2018) Extracting relationships by multi-domain matching. Adv Neural Inf Process Syst, vol 31

  19. Liu X, Guo Z, Li S et al (2021) Adversarial unsupervised domain adaptation with conditional and label shift: infer, align and iterate. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 10,367–10,376

  20. Mansour Y, Mohri M, Rostamizadeh A (2008) Domain adaptation with multiple sources. Adv Neural Inf Process Syst:21

  21. Montesuma EF, Mboula FMN (2021) Wasserstein barycenter for multi-source domain adaptation. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 16,785–16,793

  22. Peng X, Bai Q, Xia X et al (2019) Moment matching for multi-source domain adaptation. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 1406–1415

  23. Peyré G, Cuturi M, et al. (2019) Computational optimal transport: with applications to data science. Foundations Trends Mach Learn 11(5-6):355–607

    Article  MATH  Google Scholar 

  24. Podkopaev A, Ramdas A (2021) Distribution-free uncertainty quantification for classification under label shift. In: Uncertainty in artificial intelligence. PMLR, pp 844–853

  25. Rahman MM, Fookes C, Baktashmotlagh M et al (2019) Multi-component image translation for deep domain generalization. In: IEEE winter conference on applications of computer vision (WACV). IEEE, pp 579-588

  26. Rakshit S, Banerjee B, Roig G et al (2019) Unsupervised multi-source domain adaptation driven by deep adversarial ensemble learning. In: German conference on pattern recognition, Springer, pp 485–498

  27. Redko I, Courty N, Flamary R et al (2019) Optimal transport for multi-source domain adaptation under target shift. In: The 22nd international conference on artificial intelligence and statistics. PMLR, pp 849–858

  28. Redko I, Habrard A, Sebban M (2019) On the analysis of adaptability in multi-source domain adaptation. Mach Learn 108(8):1635–1652

    Article  MathSciNet  MATH  Google Scholar 

  29. Russo P, Tommasi T, Caputo B (2019) Towards multi-source adaptive semantic segmentation. In: International conference on image analysis and processing. Springer, pp 292–301

  30. Saito K, Watanabe K, Ushiku Y et al (2018) Maximum classifier discrepancy for unsupervised domain adaptation. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 3723–3732

  31. Sun B, Feng J, Saenko K (2017) Correlation alignment for unsupervised domain adaptation. In: Domain adaptation in computer vision applications. Springer, pp 153–171

  32. Sun S, Shi H, Wu Y (2015) A survey of multi-source domain adaptation. Inf Fusion 24:84–92

    Article  Google Scholar 

  33. Turrisi R, Flamary R, Rakotomamonjy A et al (2020) Multi-source domain adaptation via weighted joint distributions optimal transport. arXiv:200612938

  34. Wang H, Xu M, Ni B et al (2020) Learning to combine: knowledge aggregation for multi-source domain adaptation. In: European conference on computer vision. Springer, pp 727– 744

  35. Wang M, Deng W (2018) Deep visual domain adaptation: a survey. Neurocomputing 312:135–153

    Article  Google Scholar 

  36. Wang Z, Jing B, Ni Y, et al. (2020) Adversarial domain adaptation being aware of class relationships. In: ECAI 2020. IOS Press, Santiago de Compostela, pp 1579-1586

  37. Wilson G, Cook DJ (2020) A survey of unsupervised deep domain adaptation. ACM Trans Intell Syst Technol (TIST) 11(5): 1–46

    Article  Google Scholar 

  38. Wu H, Yan Y, Ng MK, et al. (2020) Domain-attention conditional wasserstein distance for multi-source domain adaptation. ACM Trans Intell Syst Technol (TIST) 11(4):1–19

    Article  Google Scholar 

  39. Wu H, Yan Y, Ye Y, et al. (2020) Geometric knowledge embedding for unsupervised domain adaptation. Knowl-Based Syst 191:105,155

    Article  Google Scholar 

  40. Xu R, Chen Z, Zuo W et al (2018) Deep cocktail network: multi-source unsupervised domain adaptation with category shift. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 3964–3973

  41. Yang L, Balaji Y, Lim SN et al (2020) Curriculum manager for source selection in multi-source domain adaptation. In: Computer vision–ECCV 2020: 16th European conference, Glasgow, UK, 23–28 August, 2020, proceedings, Part XIV 16. Springer, pp 608–624

  42. Zhang Y, Davison BD (2020) Impact of imagenet model selection on domain adaptation. In: Proceedings of the IEEE/CVF winter conference on applications of computer vision workshops, pp 173–182

  43. Zhao H, Zhang S, Wu G, et al. (2018) Adversarial multiple source domain adaptation. Adv Neural inf process syst 31:8559– 8570

    Google Scholar 

  44. Zhao S, Li B, Yue X, et al. (2019) Multi-source domain adaptation for semantic segmentation. Adv Neural Inf Process Syst, vol 32

  45. Zhao S, Wang G, Zhang S et al (2020) Multi-source distilling domain adaptation. In: Proceedings of the AAAI conference on artificial intelligence, pp 12,975–12,983

  46. Zhao S, Li B, Xu P, et al. (2021) Madan: multi-source adversarial domain aggregation network for domain adaptation. Int J Comput Vis:1–26

  47. Zhu Y, Zhuang F, Wang D (2019) Aligning domain-specific distribution and classifier for cross-domain classification from multiple sources. In: Proceedings of the AAAI conference on artificial intelligence, pp 5989–5996

  48. Zhuang F, Qi Z, Duan K, et al. (2020) A comprehensive survey on transfer learning. Proc IEEE 109(1):43–76

    Article  Google Scholar 

  49. Zuo Y, Yao H, Xu C (2021) Attention-based multi-source domain adaptation. IEEE Trans Image Process 30:3793–3803

    Article  MathSciNet  Google Scholar 

Download references

Acknowledgements

This work is supported by the Innovation Capacity Construction Project of Jilin Province Development and Reform Commission(2021FGWCXNLJSSZ10), the National Key Research and Development Program of China (No. 2020YFA0714103) and the Science & Technology Development Project of Jilin Province, China (20190302117GX), the Fundamental Research Funds for the Central Universities, JLU..

Author information

Authors and Affiliations

  1. College of Software, Jilin University, Changchun, 130012, Jilin, China

    Qi Wang & Shengsheng Wang

  2. Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun, 130012, Jilin, China

    Qi Wang, Shengsheng Wang & Bilin Wang

  3. College of Computer Science and Technology, Jilin University, Changchun, 130012, Jilin, China

    Shengsheng Wang & Bilin Wang

Authors
  1. Qi Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Shengsheng Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Bilin Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Shengsheng Wang.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wang, Q., Wang, S. & Wang, B. Class-rebalanced wasserstein distance for multi-source domain adaptation. Appl Intell 53, 8024–8038 (2023). https://doi.org/10.1007/s10489-022-03810-y

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10489-022-03810-y

Keywords

Search

Navigation