# Rethinking Distributional Matching Based Domain Adaptation

arXiv: https://arxiv.org/abs/2006.13352 · HTML: https://mlsanity.com/paper/2006.13352/

## Key facts

- Cited by 33 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Jun 2020.

## Built on by (most cited first, 30 of 33)

- [Cycle Self-Training for Domain Adaptation](https://mlsanity.com/paper/2103.03571/) (2021)
- [Entropy is not Enough for Test-Time Adaptation: From the Perspective of Disentangled Factors](https://mlsanity.com/paper/2403.07366/) (2024)
- [Invariant Information Bottleneck for Domain Generalization](https://mlsanity.com/paper/2106.06333/) (2021)
- [TVT: Transferable Vision Transformer for Unsupervised Domain Adaptation](https://mlsanity.com/paper/2108.05988/) (2021)
- [A Theory of Label Propagation for Subpopulation Shift](https://mlsanity.com/paper/2102.11203/) (2021)
- [Parameter-Efficient Fine-Tuning of LLaMA for the Clinical Domain](https://mlsanity.com/paper/2307.03042/) (2023)
- [Large Language Models as Annotators: Enhancing Generalization of NLP Models at Minimal Cost](https://mlsanity.com/paper/2306.15766/) (2023)
- [UMAD: Universal Model Adaptation under Domain and Category Shift](https://mlsanity.com/paper/2112.08553/) (2021)
- [Transferring Fairness under Distribution Shifts via Fair Consistency Regularization](https://mlsanity.com/paper/2206.12796/) (2022)
- [Gradual Source Domain Expansion for Unsupervised Domain Adaptation](https://mlsanity.com/paper/2311.09599/) (2023)
- [Learning CNN on ViT: A Hybrid Model to Explicitly Class-specific Boundaries for Domain Adaptation](https://mlsanity.com/paper/2403.18360/) (2024)
- [AUGCO: Augmentation Consistency-guided Self-training for Source-free Domain Adaptive Semantic Segmentation](https://mlsanity.com/paper/2107.10140/) (2021)
- [Partial Domain Adaptation without Domain Alignment](https://mlsanity.com/paper/2108.12867/) (2021)
- [Beyond Entropy: Region Confidence Proxy for Wild Test-Time Adaptation](https://mlsanity.com/paper/2505.20704/) (2025)
- [Boosting Cross-Domain Speech Recognition with Self-Supervision](https://mlsanity.com/paper/2206.09783/) (2022)
- [Are demographically invariant models and representations in medical imaging fair?](https://mlsanity.com/paper/2305.01397/) (2023)
- [Adversarial Support Alignment](https://mlsanity.com/paper/2203.08908/) (2022)
- [Domain Adaptation with Adversarial Training on Penultimate Activations](https://mlsanity.com/paper/2208.12853/) (2022)
- [De-Confusing Pseudo-Labels in Source-Free Domain Adaptation](https://mlsanity.com/paper/2401.01650/) (2024)
- [Local Context-Aware Active Domain Adaptation](https://mlsanity.com/paper/2208.12856/) (2022)
- [Encoding Hierarchical Information in Neural Networks helps in Subpopulation Shift](https://mlsanity.com/paper/2112.10844/) (2021)
- [CASUAL: Conditional Support Alignment for Domain Adaptation with Label Shift](https://mlsanity.com/paper/2305.18458/) (2023)
- [Single Image Test-Time Adaptation for Segmentation](https://mlsanity.com/paper/2309.14052/) (2023)
- [Universal Semi-Supervised Domain Adaptation by Mitigating Common-Class Bias](https://mlsanity.com/paper/2403.11234/) (2024)
- [Feature-enhanced Adversarial Semi-supervised Semantic Segmentation Network for Pulmonary Embolism Annotation](https://mlsanity.com/paper/2204.04217/) (2022)
- [Reconciling a Centroid-Hypothesis Conflict in Source-Free Domain Adaptation](https://mlsanity.com/paper/2212.03795/) (2022)
- [Contrastive Conditional Alignment based on Label Shift Calibration for Imbalanced Domain Adaptation](https://mlsanity.com/paper/2412.20337/) (2024)
- [A Simple Information-Based Approach to Unsupervised Domain-Adaptive Aspect-Based Sentiment Analysis](https://mlsanity.com/paper/2201.12549/) (2022)
- [AdaTriplet-RA: Domain Matching via Adaptive Triplet and Reinforced Attention for Unsupervised Domain Adaptation](https://mlsanity.com/paper/2211.08894/) (2022)
- [SCONE-GAN: Semantic Contrastive learning-based Generative Adversarial Network for an end-to-end image translation](https://mlsanity.com/paper/2311.03866/) (2023)

## Abstract (arXiv)

Domain adaptation (DA) is a technique that transfers predictive models trained on a labeled source domain to an unlabeled target domain, with the core difficulty of resolving distributional shift between domains. Currently, most popular DA algorithms are based on distributional matching (DM). However in practice, realistic domain shifts (RDS) may violate their basic assumptions and as a result these methods will fail. In this paper, in order to devise robust DA algorithms, we first systematically analyze the limitations of DM based methods, and then build new benchmarks with more realistic domain shifts to evaluate the well-accepted DM methods. We further propose InstaPBM, a novel Instance-based Predictive Behavior Matching method for robust DA. Extensive experiments on both conventional and RDS benchmarks demonstrate both the limitations of DM methods and the efficacy of InstaPBM: Compared with the best baselines, InstaPBM improves the classification accuracy respectively by 4.5%, 3.9% on Digits5, VisDA2017, and 2.2%, 2.9%, 3.6% on DomainNet-LDS, DomainNet-ILDS, ID-TwO. We hope our intuitive yet effective method will serve as a useful new direction and increase the robustness of DA in real scenarios. Code will be available at anonymous link: https://github.com/pikachusocute/InstaPBM-RobustDA.