mlsanityData to 7 Oct 2026

Tackling Dimensional Collapse toward Comprehensive Universal Domain Adaptation

arXiv 2410.11271 ↗First posted 15 Oct 2024cs.LGcited by 1

Not on the map. Why?

Abstract (arXiv)

Universal Domain Adaptation (UniDA) addresses unsupervised domain adaptation where target classes may differ arbitrarily from source ones, except for a shared subset. A widely used approach, partial domain matching (PDM), aligns only shared classes but struggles in extreme cases where many source classes are absent in the target domain, underperforming the most naive baseline that trains on only source data. In this work, we identify that the failure of PDM for extreme UniDA stems from dimensional collapse (DC) in target representations. To address target DC, we propose to use the de-collapse techniques in self-supervised learning on the unlabeled target data to preserve the intrinsic structure of the learned representations. Our experimental results confirm that SSL consistently advances PDM and delivers new state-of-the-art results across a broader benchmark of UniDA scenarios with different portions of shared classes, representing a crucial step toward truly comprehensive UniDA. Project page: https://dc-unida.github.io/

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 43 of its 61 references that are in the map. First posted on arXiv 15 Oct 2024.

Builds on

43 of 61 references.

References · 43

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  3. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  4. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,893
  5. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,712
  6. 2015Domain-Adversarial Training of Neural Networkscited by 3,023
  7. 2020Exploring Simple Siamese Representation Learningcited by 2,853
  8. 2021Barlow Twins: Self-Supervised Learning via Redundancy Reductioncited by 1,818
  9. 2020Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hyperspherecited by 1,393
  10. 2018Moment Matching for Multi-Source Domain Adaptationcited by 1,327
  11. 2017Adversarial Discriminative Domain Adaptationcited by 1,309
  12. 2020Energy-based Out-of-distribution Detectioncited by 1,296
31 more
  1. 2017Deep Hashing Network for Unsupervised Domain Adaptationcited by 1,181
  2. 2021VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learningcited by 1,013
  3. 2015Optimal Transport for Domain Adaptationcited by 457
  4. 2017VisDA: The Visual Domain Adaptation Challengecited by 432
  5. 2020Understanding the Behaviour of Contrastive Losscited by 397
  6. 2021Understanding Dimensional Collapse in Contrastive Self-supervised Learningcited by 374
  7. 2017Central Moment Discrepancy (CMD) for Domain-Invariant Representation Learningcited by 192
  8. 2020Universal Domain Adaptation through Self Supervisioncited by 172
  9. 2018Partial Adversarial Domain Adaptationcited by 127
  10. 2021Self-supervised Learning is More Robust to Dataset Imbalancecited by 122
  11. 2019Unsupervised Domain Adaptation through Self-Supervisioncited by 116
  12. 2020Revisiting Training Strategies and Generalization Performance in Deep Metric Learningcited by 99
  13. 2020Self-Supervised Learning for Domain Adaptation on Point-Cloudscited by 99
  14. 2021OVANet: One-vs-All Network for Universal Domain Adaptationcited by 92
  15. 2020Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptationcited by 60
  16. 2019Self-supervised Domain Adaptation for Computer Vision Taskscited by 58
  17. 2022Unified Optimal Transport Framework for Universal Domain Adaptationcited by 57
  18. 2021Mimicking the Oracle: An Initial Phase Decorrelation Approach for Class Incremental Learningcited by 49
  19. 2022Towards Understanding and Mitigating Dimensional Collapse in Heterogeneous Federated Learningcited by 45
  20. 2020Self-Supervised Learning Across Domainscited by 44
  21. 2024LEAD: Learning Decomposition for Source-free Universal Domain Adaptationcited by 25
  22. 2020Extending and Analyzing Self-Supervised Learning Across Domainscited by 20
  23. 2022Subsidiary Prototype Alignment for Universal Domain Adaptationcited by 17
  24. 2023Universal Domain Adaptation via Compressive Attention Matchingcited by 15
  25. 2023MLNet: Mutual Learning Network with Neighborhood Invariance for Universal Domain Adaptationcited by 14
  26. 2023Universal Domain Adaptation from Foundation Models: A Baseline Studycited by 10
  27. 2019Tackling Partial Domain Adaptation with Self-Supervisioncited by 9
  28. 2021Can Pretext-Based Self-Supervised Learning Be Boosted by Downstream Data? A Theoretical Analysiscited by 9
  29. 2022Learning Classifiers of Prototypes and Reciprocal Points for Universal Domain Adaptationcited by 8
  30. 2023Generalized Universal Domain Adaptation with Generative Flow Networkscited by 7
  31. 2023Complementary Benefits of Contrastive Learning and Self-Training Under Distribution Shiftcited by 5

Built on by

Cited by year

Citing papers per year: 2025: 1.

1 paper.

  1. 2025Domain Adaptation via Feature Refinementcited by 0