pith. sign in

arxiv: 2501.17443 · v4 · pith:IMRUATK2new · submitted 2025-01-29 · 💻 cs.LG

Gradual Domain Adaptation for Graph Learning

Pui Ieng Lei , Ximing Chen , Yijun Sheng , Yanyan Liu , Zhiguo Gong , Qiang Yang This is my paper

Pith reviewed 2026-05-23 04:39 UTC · model grok-4.3

classification 💻 cs.LG
keywords gradual domain adaptationgraph domain adaptationFused Gromov-WassersteinWasserstein distancevertex-based progressiontransfer learninggraph neural networks
0
0 comments X

The pith

A graph gradual domain adaptation framework constructs sequences of intermediate graphs over the Fused Gromov-Wasserstein metric and supplies bounds on consecutive Wasserstein distances to enable adaptation across large shifts.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents the GGDA framework to address the absence of graph domain adaptation methods that can manage large distribution shifts between source and target graphs. It generates a pool of knowledge-preserving intermediate graphs using the Fused Gromov-Wasserstein metric and then builds a compact domain sequence through vertex-based progression that selects nearby vertices for adaptive advancement. The approach derives implementable upper and lower bounds on the inter-domain Wasserstein distance W_p between consecutive distributions, allowing control over domain spacing to reduce information loss. Experiments on diverse transfer tasks show improved performance over prior techniques that cannot simulate coherent evolutionary paths.

Core claim

The GGDA framework constructs a compact domain sequence that minimizes information loss during adaptation by first generating knowledge-preserving intermediate graphs over the Fused Gromov-Wasserstein metric and then applying vertex-based progression to select close vertices and perform adaptive domain advancement, while supplying implementable upper and lower bounds on the intractable inter-domain Wasserstein distance W_p(μ_t, μ_{t+1}) that permit its flexible adjustment for optimal domain formation.

What carries the argument

Vertex-based progression over a Fused Gromov-Wasserstein bridging data pool, which selects close vertices and advances domains adaptively while enabling explicit bounds on consecutive Wasserstein distances.

If this is right

  • The bounds enable explicit control over the number and spacing of intermediate domains to minimize total adaptation cost.
  • The vertex-based selection step produces domain sequences that preserve graph structure better than uniform interpolation.
  • The method applies to any graph transfer task where direct source-to-target shifts are too large for existing adaptation techniques.
  • Performance gains appear across multiple graph datasets and shift magnitudes in the reported experiments.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same progression technique could be tested on dynamic or temporal graphs to track evolving structures over time.
  • If the FGW generation step scales, the framework might reduce reliance on target labels in semi-supervised graph settings.
  • The distance bounds could be reused in other optimal-transport domain adaptation pipelines that currently treat inter-domain distances as black boxes.
  • An open question left by the work is whether the same bounds extend to non-graph structured data when analogous metrics are substituted.

Load-bearing premise

Knowledge-preserving intermediate graphs can be generated efficiently over the Fused Gromov-Wasserstein metric and the vertex-based progression will improve transferability without adding biases or losing information.

What would settle it

A dataset where the computed bounds on W_p(μ_t, μ_{t+1}) fail to predict actual transfer accuracy or where the generated intermediate graphs produce lower accuracy than direct adaptation would falsify the central claim.

Figures

Figures reproduced from arXiv: 2501.17443 by Pui Ieng Lei, Qiang Yang, Ximing Chen, Yanyan Liu, Yijun Sheng, Zhiguo Gong.

Figure 1
Figure 1. Figure 1: Illustration of our GGDA framework’s design rationale. The upper part shows the sequence of intermediate graphs generated [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Node classification accuracy on target graphs with varying levels of discrepancy from the source graph. The x-axis shows the [PITH_FULL_IMAGE:figures/full_fig_p017_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Progression of the constructed domain sequence under different [PITH_FULL_IMAGE:figures/full_fig_p018_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Performance of our GGDA framework with varying values of hyperparameters [PITH_FULL_IMAGE:figures/full_fig_p018_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Visualization of mass decay in each stage of GGDA. The x-axis represents the set of the top-weighted vertices constituting [PITH_FULL_IMAGE:figures/full_fig_p019_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Visualization of target graph node embeddings for the task A2 to D2, where distinct colors signify different class labels. The [PITH_FULL_IMAGE:figures/full_fig_p019_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Each plane represents a 2-dimensional space [PITH_FULL_IMAGE:figures/full_fig_p020_7.png] view at source ↗
read the original abstract

Existing machine learning literature lacks graph-based domain adaptation techniques capable of handling large distribution shifts, primarily due to the difficulty in simulating a coherent evolutionary path from source to target graph. To meet this challenge, we present a graph gradual domain adaptation (GGDA) framework, which constructs a compact domain sequence that minimizes information loss during adaptation. Our approach starts with an efficient generation of knowledge-preserving intermediate graphs over the Fused Gromov-Wasserstein (FGW) metric. A GGDA domain sequence is then constructed upon this bridging data pool through a novel vertex-based progression, which involves selecting "close" vertices and performing adaptive domain advancement to enhance inter-domain transferability. Theoretically, our framework provides implementable upper and lower bounds for the intractable inter-domain Wasserstein distance, $W_p(\mu_t,\mu_{t+1})$, enabling its flexible adjustment for optimal domain formation. Extensive experiments across diverse transfer scenarios demonstrate the superior performance of our GGDA framework.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces a graph gradual domain adaptation (GGDA) framework to handle large distribution shifts in graph learning. It generates knowledge-preserving intermediate graphs via the Fused Gromov-Wasserstein (FGW) metric, constructs a domain sequence using a novel vertex-based progression with adaptive domain advancement, and claims to supply implementable upper and lower bounds on the inter-domain Wasserstein distance W_p(μ_t, μ_{t+1}) to enable optimal domain formation. Extensive experiments across transfer scenarios are reported to show superior performance over existing methods.

Significance. If the claimed bounds on W_p(μ_t, μ_{t+1}) can be evaluated without incurring the cost of a fresh optimal transport solve and the vertex-based progression demonstrably improves transferability without new biases, the framework would address a recognized gap in graph domain adaptation for large shifts. The use of FGW for bridging graphs and the provision of explicit bounds (if implementable) would constitute a concrete technical contribution.

major comments (2)
  1. [Abstract / theoretical analysis] Abstract and theoretical section: the central claim that the framework supplies 'implementable' upper and lower bounds for the intractable W_p(μ_t, μ_{t+1}) is load-bearing, yet no explicit form, derivation, or complexity analysis is visible showing that these bounds are obtained from precomputed FGW quantities or closed-form expressions rather than auxiliary OT problems whose cost matches the original intractability. This directly affects whether the bounds enable 'flexible adjustment for optimal domain formation' without circularity or prohibitive computation.
  2. [Method / vertex-based progression] Method section on vertex-based progression: the claim that selecting 'close' vertices and performing adaptive domain advancement enhances inter-domain transferability rests on the assumption that this process avoids introducing new biases or information loss; no quantitative verification (e.g., via information-theoretic measures or ablation on progression steps) is referenced to support that the progression is strictly knowledge-preserving.
minor comments (2)
  1. [Introduction] Notation for the domain sequence {μ_t} and the precise definition of the FGW-based bridging pool should be introduced earlier and used consistently to improve readability.
  2. [Experiments] Experimental section: tables reporting performance should include standard deviations over multiple runs and explicit baseline implementations to allow direct comparison of the reported gains.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We address each major point below and will incorporate clarifications and additional analyses in the revised manuscript.

read point-by-point responses

  1. Referee: [Abstract / theoretical analysis] Abstract and theoretical section: the central claim that the framework supplies 'implementable' upper and lower bounds for the intractable W_p(μ_t, μ_{t+1}) is load-bearing, yet no explicit form, derivation, or complexity analysis is visible showing that these bounds are obtained from precomputed FGW quantities or closed-form expressions rather than auxiliary OT problems whose cost matches the original intractability. This directly affects whether the bounds enable 'flexible adjustment for optimal domain formation' without circularity or prohibitive computation.

    Authors: We agree that the explicit forms, derivations, and complexity analysis of the upper and lower bounds on W_p(μ_t, μ_{t+1}) require clearer presentation. These bounds are constructed directly from the precomputed FGW distances and vertex correspondences already obtained during intermediate graph generation, without requiring new optimal transport solves. We will expand the theoretical section with the full closed-form expressions, step-by-step derivation, and O(n^2) complexity discussion to demonstrate implementability and remove any ambiguity. revision: yes

  2. Referee: [Method / vertex-based progression] Method section on vertex-based progression: the claim that selecting 'close' vertices and performing adaptive domain advancement enhances inter-domain transferability rests on the assumption that this process avoids introducing new biases or information loss; no quantitative verification (e.g., via information-theoretic measures or ablation on progression steps) is referenced to support that the progression is strictly knowledge-preserving.

    Authors: The vertex-based progression is designed to select vertices with minimal FGW cost to preserve structure, but we acknowledge the absence of explicit quantitative checks such as mutual information retention or step-wise ablations. We will add these analyses, including information-theoretic metrics and ablation studies on the number of progression steps, to the experimental section to empirically confirm knowledge preservation. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation of implementable bounds or GGDA sequence

full rationale

The paper constructs intermediate graphs via FGW metric, then applies vertex-based progression to form the domain sequence, and separately states that this yields implementable upper/lower bounds on W_p(μ_t, μ_{t+1}). No equation or step is shown reducing the bound to a fitted parameter, self-definition, or prior self-citation that itself assumes the target result. The FGW generation and progression are independent of the bound claim; the bounds are presented as a theoretical consequence rather than tautological. No self-citation load-bearing, ansatz smuggling, or renaming of known results appears in the abstract or described chain. The framework is self-contained against the external FGW metric.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Only the abstract is available, so the ledger is necessarily incomplete and based on high-level claims.

axioms (1)
  • domain assumption The Fused Gromov-Wasserstein metric can generate a compact sequence of knowledge-preserving intermediate graphs between source and target domains.
    Invoked as the starting point for constructing the bridging data pool.

pith-pipeline@v0.9.0 · 5700 in / 1163 out tokens · 28378 ms · 2026-05-23T04:39:20.521810+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

85 extracted references · 85 canonical work pages · 3 internal anchors

  1. [1]

    Samira Abnar, Rianne van den Berg, Golnaz Ghiasi, Mostafa Dehghani, Nal Kalchbrenner, and Hanie Sedghi. 2021. Gradual domain adaptation in the wild: When intermediate distributions are absent. arXiv preprint arXiv:2106.06080 (2021)

    work page arXiv 2021

  2. [2]

    Konstantinos Bousmalis, George Trigeorgis, Nathan Silberman, Dilip Krishnan, and Dumitru Erhan. 2016. Domain separation networks. Advances in neural information processing systems 29 (2016)

    work page 2016

  3. [3]

    Ruichu Cai, Fengzhu Wu, Zijian Li, Pengfei Wei, Lingling Yi, and Kun Zhang. 2024. Graph domain adaptation: A generative view. ACM Transactions on Knowledge Discovery from Data 18, 3 (2024), 1–24

    work page 2024

  4. [4]

    Benson Chen, Gary Bécigneul, Octavian-Eugen Ganea, Regina Barzilay, and Tommi Jaakkola. 2020. Optimal transport graph neural networks. arXiv preprint arXiv:2006.04804 (2020)

    work page arXiv 2020

  5. [5]

    Hong-You Chen and Wei-Lun Chao. 2021. Gradual domain adaptation without indexed intermediate domains. Advances in neural information processing systems 34 (2021), 8201–8214

    work page 2021

  6. [6]

    Wei Chen, Guo Ye, Yakun Wang, Zhao Zhang, Libang Zhang, Daixin Wang, Zhiqiang Zhang, and Fuzhen Zhuang. 2025. Smoothness Really Matters: A Simple Yet Effective Approach for Unsupervised Graph Domain Adaptation. InProceedings of the AAAI Conference on Artificial Intelligence , Vol. 39. 15875–15883

    work page 2025

  7. [7]

    Samir Chowdhury and Tom Needham. 2021. Generalized spectral clustering via Gromov-Wasserstein learning. In International Conference on Artificial Intelligence and Statistics. PMLR, 712–720

    work page 2021

  8. [8]

    Nicolas Courty, Rémi Flamary, Devis Tuia, and Alain Rakotomamonjy. 2016. Optimal transport for domain adaptation. IEEE transactions on pattern analysis and machine intelligence 39, 9 (2016), 1853–1865

    work page 2016

  9. [9]

    Quanyu Dai, Xiao-Ming Wu, Jiaren Xiao, Xiao Shen, and Dan Wang. 2022. Graph transfer learning via adversarial domain adaptation with graph convolution. IEEE Transactions on Knowledge and Data Engineering 35, 5 (2022), 4908–4922

    work page 2022

  10. [10]

    Michaël Defferrard, Xavier Bresson, and Pierre Vandergheynst. 2016. Convolutional neural networks on graphs with fast localized spectral filtering. Advances in neural information processing systems 29 (2016)

    work page 2016

  11. [11]

    Yash Deshpande, Subhabrata Sen, Andrea Montanari, and Elchanan Mossel. 2018. Contextual stochastic block models.Advances in Neural Information Processing Systems 31 (2018)

    work page 2018

  12. [12]

    Cornelia Druţu and Michael Kapovich. 2018. Geometric group theory. Vol. 63. American Mathematical Soc

    work page 2018

  13. [13]

    Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pascal Germain, Hugo Larochelle, François Laviolette, Mario March, and Victor Lempitsky. 2016. Domain-adversarial training of neural networks. Journal of machine learning research 17, 59 (2016), 1–35

    work page 2016

  14. [14]

    Vikas Garg, Stefanie Jegelka, and Tommi Jaakkola. 2020. Generalization and representational limits of graph neural networks. In International Conference on Machine Learning . PMLR, 3419–3430

    work page 2020

  15. [15]

    Justin Gilmer, Samuel S Schoenholz, Patrick F Riley, Oriol Vinyals, and George E Dahl. 2017. Neural message passing for quantum chemistry. In International conference on machine learning . PMLR, 1263–1272

    work page 2017

  16. [16]

    Boqing Gong, Yuan Shi, Fei Sha, and Kristen Grauman. 2012. Geodesic flow kernel for unsupervised domain adaptation. In 2012 IEEE conference on computer vision and pattern recognition . IEEE, 2066–2073

    work page 2012

  17. [17]

    Rui Gong, Wen Li, Yuhua Chen, and Luc Van Gool. 2019. Dlow: Domain flow for adaptation and generalization. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 2477–2486

    work page 2019

  18. [18]

    Raghuraman Gopalan, Ruonan Li, and Rama Chellappa. 2011. Domain adaptation for object recognition: An unsupervised approach. In 2011 international conference on computer vision . IEEE, 999–1006

    work page 2011

  19. [19]

    Gaoyang Guo, Chaokun Wang, Bencheng Yan, Yunkai Lou, Hao Feng, Junchao Zhu, Jun Chen, Fei He, and S Yu Philip. 2022. Learning adaptive node embeddings across graphs. IEEE Transactions on Knowledge and Data Engineering 35, 6 (2022), 6028–6042

    work page 2022

  20. [20]

    Will Hamilton, Zhitao Ying, and Jure Leskovec. 2017. Inductive representation learning on large graphs. Advances in neural information processing systems 30 (2017)

    work page 2017

  21. [21]

    Yifei He, Haoxiang Wang, Bo Li, and Han Zhao. 2023. Gradual Domain Adaptation: Theory and Algorithms. arXiv preprint arXiv:2310.13852 (2023)

    work page arXiv 2023

  22. [22]

    Han-Kai Hsu, Chun-Han Yao, Yi-Hsuan Tsai, Wei-Chih Hung, Hung-Yu Tseng, Maneesh Singh, and Ming-Hsuan Yang. 2020. Progressive domain adaptation for object detection. In Proceedings of the IEEE/CVF winter conference on applications of computer vision . 749–757

    work page 2020

  23. [23]

    Weihua Hu, Matthias Fey, Marinka Zitnik, Yuxiao Dong, Hongyu Ren, Bowen Liu, Michele Catasta, and Jure Leskovec. 2020. Open graph benchmark: Datasets for machine learning on graphs. Advances in neural information processing systems 33 (2020), 22118–22133

    work page 2020

  24. [24]

    Renhong Huang, Jiarong Xu, Xin Jiang, Ruichuan An, and Yang Yang. 2024. Can Modifying Data Address Graph Domain Adaptation?. InProceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 1131–1142

    work page 2024

  25. [25]

    Leonid V Kantorovich. 1960. Mathematical methods of organizing and planning production. Management science 6, 4 (1960), 366–422

    work page 1960

  26. [26]

    George Karypis and Vipin Kumar. 1998. A fast and high quality multilevel scheme for partitioning irregular graphs. SIAM Journal on scientific Computing 20, 1 (1998), 359–392

    work page 1998

  27. [27]

    Thomas N Kipf and Max Welling. 2016. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907 (2016)

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  28. [28]

    Ananya Kumar, Tengyu Ma, and Percy Liang. 2020. Understanding self-training for gradual domain adaptation. In International conference on machine learning. PMLR, 5468–5479. Manuscript submitted to ACM Gradual Domain Adaptation for Graph Learning 23

    work page 2020

  29. [29]

    Jundong Li, Xia Hu, Jiliang Tang, and Huan Liu. 2015. Unsupervised streaming feature selection in social media. In Proceedings of the 24th ACM International on Conference on Information and Knowledge Management . 1041–1050

    work page 2015

  30. [30]

    Jiajin Li, Jianheng Tang, Lemin Kong, Huikang Liu, Jia Li, Anthony Man-Cho So, and Jose Blanchet. 2023. A convergent single-loop algorithm for relaxation of gromov-wasserstein in graph data. arXiv preprint arXiv:2303.06595 (2023)

    work page arXiv 2023

  31. [31]

    Meihan Liu, Zeyu Fang, Zhen Zhang, Ming Gu, Sheng Zhou, Xin Wang, and Jiajun Bu. 2024. Rethinking Propagation for Unsupervised Graph Domain Adaptation. arXiv preprint arXiv:2402.05660 (2024)

    work page arXiv 2024

  32. [32]

    Ming-Yu Liu and Oncel Tuzel. 2016. Coupled generative adversarial networks. Advances in neural information processing systems 29 (2016)

    work page 2016

  33. [33]

    Shikun Liu, Tianchun Li, Yongbin Feng, Nhan Tran, Han Zhao, Qiang Qiu, and Pan Li. 2023. Structural re-weighting improves graph domain adaptation. In International Conference on Machine Learning . PMLR, 21778–21793

    work page 2023

  34. [34]

    Shikun Liu, Deyu Zou, Han Zhao, and Pan Li. 2024. Pairwise Alignment Improves Graph Domain Adaptation.arXiv preprint arXiv:2403.01092 (2024)

    work page arXiv 2024

  35. [35]

    Mingsheng Long, Zhangjie Cao, Jianmin Wang, and Michael I Jordan. 2018. Conditional adversarial domain adaptation. Advances in neural information processing systems 31 (2018)

    work page 2018

  36. [36]

    Mingsheng Long, Han Zhu, Jianmin Wang, and Michael I Jordan. 2017. Deep transfer learning with joint adaptation networks. In International conference on machine learning . PMLR, 2208–2217

    work page 2017

  37. [37]

    Xinyu Ma, Xu Chu, Yasha Wang, Yang Lin, Junfeng Zhao, Liantao Ma, and Wenwu Zhu. 2024. Fused Gromov-Wasserstein Graph Mixup for Graph-level Classifications. Advances in Neural Information Processing Systems 36 (2024)

    work page 2024

  38. [38]

    Yuzhen Mao, Jianhui Sun, and Dawei Zhou. 2022. Augmenting Knowledge Transfer across Graphs. In 2022 IEEE International Conference on Data Mining (ICDM). IEEE, 1101–1106

    work page 2022

  39. [39]

    Facundo Mémoli. 2011. Gromov–Wasserstein distances and the metric approach to object matching. Foundations of computational mathematics 11 (2011), 417–487

    work page 2011

  40. [40]

    Jaemin Na, Heechul Jung, Hyung Jin Chang, and Wonjun Hwang. 2021. Fixbi: Bridging domain spaces for unsupervised domain adaptation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 1094–1103

    work page 2021

  41. [41]

    Sinno Jialin Pan and Qiang Yang. 2009. A survey on transfer learning. IEEE Transactions on knowledge and data engineering 22, 10 (2009), 1345–1359

    work page 2009

  42. [42]

    Gabriel Peyré, Marco Cuturi, et al. 2019. Computational optimal transport: With applications to data science. Foundations and Trends® in Machine Learning 11, 5-6 (2019), 355–607

    work page 2019

  43. [43]

    Gabriel Peyré, Marco Cuturi, and Justin Solomon. 2016. Gromov-wasserstein averaging of kernel and distance matrices. In International conference on machine learning. PMLR, 2664–2672

    work page 2016

  44. [44]

    Ziyue Qiao, Xiao Luo, Meng Xiao, Hao Dong, Yuanchun Zhou, and Hui Xiong. 2023. Semi-supervised domain adaptation in graph transfer learning. arXiv preprint arXiv:2309.10773 (2023)

    work page arXiv 2023

  45. [45]

    Benedek Rozemberczki, Carl Allen, and Rik Sarkar. 2021. Multi-scale attributed node embedding. Journal of Complex Networks 9, 2 (2021), cnab014

    work page 2021

  46. [46]

    Yossi Rubner, Carlo Tomasi, and Leonidas J Guibas. 2000. The earth mover’s distance as a metric for image retrieval.International journal of computer vision 40 (2000), 99–121

    work page 2000

  47. [47]

    Shogo Sagawa and Hideitsu Hino. 2022. Gradual domain adaptation via normalizing flows. arXiv preprint arXiv:2206.11492 (2022)

    work page arXiv 2022

  48. [48]

    Meyer Scetbon, Gabriel Peyré, and Marco Cuturi. 2022. Linear-time gromov wasserstein distances using low rank couplings and costs. InInternational Conference on Machine Learning . PMLR, 19347–19365

    work page 2022

  49. [49]

    Prithviraj Sen, Galileo Namata, Mustafa Bilgic, Lise Getoor, Brian Galligher, and Tina Eliassi-Rad. 2008. Collective classification in network data. AI magazine 29, 3 (2008), 93–93

    work page 2008

  50. [50]

    Kendrick Shen, Robbie M Jones, Ananya Kumar, Sang Michael Xie, Jeff Z HaoChen, Tengyu Ma, and Percy Liang. 2022. Connect, not collapse: Explaining contrastive learning for unsupervised domain adaptation. In International conference on machine learning . PMLR, 19847–19878

    work page 2022

  51. [51]

    Xiao Shen, Quanyu Dai, Fu-lai Chung, Wei Lu, and Kup-Sze Choi. 2020. Adversarial deep network embedding for cross-network node classification. In Proceedings of the AAAI conference on artificial intelligence , Vol. 34. 2991–2999

    work page 2020

  52. [52]

    Boshen Shi, Yongqing Wang, Fangda Guo, Jiangli Shao, Huawei Shen, and Xueqi Cheng. 2023. Improving graph domain adaptation with network hierarchy. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management . 2249–2258

    work page 2023

  53. [53]

    Lianghe Shi and Weiwei Liu. 2024. Adversarial Self-Training Improves Robustness and Generalization for Gradual Domain Adaptation. Advances in Neural Information Processing Systems 36 (2024)

    work page 2024

  54. [54]

    Hidetoshi Shimodaira. 2000. Improving predictive inference under covariate shift by weighting the log-likelihood function. Journal of statistical planning and inference 90, 2 (2000), 227–244

    work page 2000

  55. [55]

    Ankit Singh. 2021. Clda: Contrastive learning for semi-supervised domain adaptation. Advances in Neural Information Processing Systems 34 (2021), 5089–5101

    work page 2021

  56. [56]

    Ben Tan, Yangqiu Song, Erheng Zhong, and Qiang Yang. 2015. Transitive transfer learning. In Proceedings of the 21th ACM SIGKDD international conference on knowledge discovery and data mining . 1155–1164

    work page 2015

  57. [57]

    Ben Tan, Yu Zhang, Sinno Pan, and Qiang Yang. 2017. Distant domain transfer learning. InProceedings of the AAAI conference on artificial intelligence , Vol. 31

    work page 2017

  58. [58]

    Vayer Titouan, Nicolas Courty, Romain Tavenard, and Rémi Flamary. 2019. Optimal transport for structured data with application on graphs. In International Conference on Machine Learning . PMLR, 6275–6284. Manuscript submitted to ACM 24 Lei et al

    work page 2019

  59. [59]

    Eric Tzeng, Judy Hoffman, Kate Saenko, and Trevor Darrell. 2017. Adversarial discriminative domain adaptation. InProceedings of the IEEE conference on computer vision and pattern recognition . 7167–7176

    work page 2017

  60. [60]

    Laurens Van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-SNE. Journal of machine learning research 9, 11 (2008)

    work page 2008

  61. [61]

    Titouan Vayer, Laetitia Chapel, Rémi Flamary, Romain Tavenard, and Nicolas Courty. 2020. Fused Gromov-Wasserstein distance for structured objects. Algorithms 13, 9 (2020), 212

    work page 2020

  62. [62]

    Petar Veličković, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Lio, and Yoshua Bengio. 2017. Graph attention networks.arXiv preprint arXiv:1710.10903 (2017)

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  63. [63]

    Cédric Vincent-Cuaz, Rémi Flamary, Marco Corneli, Titouan Vayer, and Nicolas Courty. 2021. Semi-relaxed Gromov-Wasserstein divergence with applications on graphs. arXiv preprint arXiv:2110.02753 (2021)

    work page arXiv 2021

  64. [64]

    Cédric Vincent-Cuaz, Rémi Flamary, Marco Corneli, Titouan Vayer, and Nicolas Courty. 2022. Template based graph neural network with optimal transport distances. Advances in Neural Information Processing Systems 35 (2022), 11800–11814

    work page 2022

  65. [65]

    Cédric Vincent-Cuaz, Titouan Vayer, Rémi Flamary, Marco Corneli, and Nicolas Courty. 2021. Online graph dictionary learning. InInternational conference on machine learning . PMLR, 10564–10574

    work page 2021

  66. [66]

    Haoxiang Wang, Bo Li, and Han Zhao. 2022. Understanding gradual domain adaptation: Improved analysis, optimal path and beyond. InInternational Conference on Machine Learning . PMLR, 22784–22801

    work page 2022

  67. [67]

    Yu Wang, Ronghang Zhu, Pengsheng Ji, and Sheng Li. 2024. Open-Set Graph Domain Adaptation via Separate Domain Alignment. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 38. 9142–9150

    work page 2024

  68. [68]

    Jun Wu, Jingrui He, and Elizabeth Ainsworth. 2023. Non-iid transfer learning on graphs. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 37. 10342–10350

    work page 2023

  69. [69]

    Man Wu, Shirui Pan, Chuan Zhou, Xiaojun Chang, and Xingquan Zhu. 2020. Unsupervised domain adaptive graph convolutional networks. In Proceedings of The Web Conference 2020 . 1457–1467

    work page 2020

  70. [70]

    Man Wu, Shirui Pan, and Xingquan Zhu. 2022. Attraction and repulsion: Unsupervised domain adaptive graph contrastive learning network. IEEE Transactions on Emerging Topics in Computational Intelligence 6, 5 (2022), 1079–1091

    work page 2022

  71. [71]

    Yifan Wu, Ezra Winston, Divyansh Kaushik, and Zachary Lipton. 2019. Domain adaptation with asymmetrically-relaxed distribution alignment. In International conference on machine learning . PMLR, 6872–6881

    work page 2019

  72. [72]

    Hongteng Xu, Dixin Luo, and Lawrence Carin. 2019. Scalable Gromov-Wasserstein learning for graph partitioning and matching. Advances in neural information processing systems 32 (2019)

    work page 2019

  73. [73]

    Hongteng Xu, Dixin Luo, Hongyuan Zha, and Lawrence Carin Duke. 2019. Gromov-wasserstein learning for graph matching and node embedding. In International conference on machine learning . PMLR, 6932–6941

    work page 2019

  74. [74]

    Keyulu Xu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. 2018. How powerful are graph neural networks?arXiv preprint arXiv:1810.00826 (2018)

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  75. [75]

    Minghao Xu, Jian Zhang, Bingbing Ni, Teng Li, Chengjie Wang, Qi Tian, and Wenjun Zhang. 2020. Adversarial domain adaptation with domain mixup. In Proceedings of the AAAI conference on artificial intelligence , Vol. 34. 6502–6509

    work page 2020

  76. [76]

    Nan Yin, Li Shen, Baopu Li, Mengzhu Wang, Xiao Luo, Chong Chen, Zhigang Luo, and Xian-Sheng Hua. 2022. Deal: An unsupervised domain adaptive framework for graph-level classification. In Proceedings of the 30th ACM International Conference on Multimedia . 3470–3479

    work page 2022

  77. [77]

    Nan Yin, Li Shen, Mengzhu Wang, Long Lan, Zeyu Ma, Chong Chen, Xian-Sheng Hua, and Xiao Luo. 2023. Coco: A coupled contrastive framework for unsupervised domain adaptive graph classification. In International Conference on Machine Learning . PMLR, 40040–40053

    work page 2023

  78. [78]

    Yuning You, Tianlong Chen, Zhangyang Wang, and Yang Shen. 2023. Graph domain adaptation via theory-grounded spectral regularization. In The eleventh international conference on learning representations

    work page 2023

  79. [79]

    Xiaowen Zhang, Yuntao Du, Rongbiao Xie, and Chongjun Wang. 2021. Adversarial separation network for cross-network node classification. In Proceedings of the 30th ACM international conference on information & knowledge management . 2618–2626

    work page 2021

  80. [80]

    Yuchen Zhang, Tianle Liu, Mingsheng Long, and Michael Jordan. 2019. Bridging theory and algorithm for domain adaptation. In International conference on machine learning . PMLR, 7404–7413

    work page 2019

Showing first 80 references.