Pith. sign in

REVIEW 5 major objections 3 minor 57 references

Smoothness Really Matters: A Simple Yet Effective Approach for Unsupervised Graph Domain Adaptation

T0 review · 5 major / 3 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read The paper proves that adding a Laplacian smoothing loss on the target graph reduces an explicit smoothness term in the target-risk bound, and shows the resulting method outperforms all baselines on six graph-transfer tasks.

desk verdict A simple, plausibly useful smoothing regularizer for UGDA with a theory section that does not prove what it claims; worth engaging on the empirics, but the theory needs major repair or a big step back. read the letter →

arxiv 2412.11654 v3 pith:W3QNSPFQ submitted 2024-12-16 cs.LG cs.AI

classification cs.LGcs.AI
keywords unsupervisedgraphdomainadaptationneuralnetworksstructuraldistributionshiftLaplaciansmoothingrandomwalksamplingmodelsmoothnesstargetriskboundtransferlearning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper takes on unsupervised graph domain adaptation: learning a classifier on a labeled source graph and applying it to an unlabeled target graph whose structure differs. The authors argue that prior methods focus on aligning feature distributions and neglect structural shifts, which hit graph neural networks especially hard because small local-structure changes can move embeddings sharply. Their central claim is a theoretical bound showing target risk is controlled by source risk, domain discrepancy, and the model's smoothness on the two graphs. On that basis they introduce TDSS, a plug-in Laplacian smoothing loss computed on the target graph with random-walk-sampled neighbors, which directly penalizes feature variation between neighboring nodes. Across six transfer tasks on three citation networks, TDSS outperforms all baselines, with average gains of 2.04% Macro-F1 and 3.05% Micro-F1 over the previous best method.

What carries the argument

The load-bearing object is the pair consisting of Theorem 1 and the smoothing loss $L_{SR}$: a target-risk bound in which model smoothness $\Phi$ appears as an additive term, and a Laplacian smoothing loss $$L_{SR} = \frac{1}{2}\sum_i\sum_j \tilde A_{ij}\left\|\frac{f(x_i)}{\sqrt{d_i}} - \frac{f(x_j)}{\sqrt{d_j}}\right\|^2$$ computed on a random-walk-sampled adjacency $\tilde A$ of the target graph. Definition 1 defines $\Phi$ as the expected supremum of output variation over nodes within graph distance $k$ and feature distance $r$; the proof ties $\Phi$ to worst-case loss under Wasserstein-$\infty$ distribution shift via a covering-number argument. The random-walk sampling in Eqs. (2)-(3) is what lets the smoothing loss control local structure without collapsing embeddings, and the overall loss Eq. (13) mirrors the three terms of the bound.

What would settle it

On any transfer task, train the identical GNN with and without the smoothing loss, then measure how much the model's predictions vary between nearby feature points on the target graph using the trained models; if the smoothing term does not reduce this measured smoothness, the claimed mechanism is falsified even if accuracy improves.

Watch

Extended reading notes

Core claim

The paper's central discovery is a new target-risk bound, Theorem 1: with high probability, $E_T(f) \le E_S(f) + 2\mathrm{TVD}(S,T) + \Phi_S + \Phi_T + K$, where $\Phi$ is the model smoothness of Definition 1 and $K$ collects finite-sample concentration terms. This bound says that lowering the smoothness constant on the target graph directly lowers the worst-case target error, independent of the source error and domain discrepancy. The paper operationalizes the bound by replacing $2\Phi_T$ with a Laplacian smoothing loss $L_{SR}$ on the target graph, with neighbors chosen by random walk to avoid over-smoothing, and aligns the other terms with classifier loss $L_{GC}$ and MMD-based domain alignment $L_{DA}$, giving $L = L_{GC}+\alpha L_{DA}+\beta L_{SR}$.

Load-bearing premise

The theory assumes that smoothing the graph's edges actually makes the model's predictions smoother across nearby feature points in the sense of Definition 1, and that the source graph is already smoother than the target graph; the paper asserts this link rather than proving it.

Editorial extensions

If this is right

  • Any GNN backbone can receive the smoothing loss unchanged; the paper reports consistent gains with GCN, GAT, and SGC as well as with the A2GNN backbone.
  • Because the bound shows target-side smoothness is a direct lever on target risk, combining smoothing with existing alignment methods is a general recipe rather than a competing approach.
  • Random-walk sampling is essential: replacing it with fixed k-hop sampling reduces gains, indicating that sampling controls over-smoothing while preserving the smoothing benefit.
  • The loss is computed only on unlabeled target nodes, so TDSS works in fully unsupervised transfer and adds only $O(n\gamma\lambda + n\rho d)$ time.
  • The average gains over A2GNN are 2.04 Macro-F1 and 3.05 Micro-F1 points across all six transfer scenarios.
  • The three-term decomposition suggests that future work can attack target risk by improving any of source accuracy, domain alignment, or target smoothness independently.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same smoothing principle could extend to source-free or test-time graph adaptation, where no source labels are available at adaptation time, because the loss needs only the target graph and the model's own outputs.
  • The asserted link between adjacency-based smoothing and the feature-space smoothness constant is testable: if it fails on heterophilic or long-range graphs, the method's gains on those settings should shrink.
  • Since Eq. (11) makes the bound metric-agnostic under kernel conditions, replacing MMD with another discrepancy measure should preserve the theory while possibly changing the empirical trade-off.
  • The assumption $\Phi_S < \Phi_T$ may be violated when transferring from a complex source to a simpler target, predicting smaller or negative gains; checking TDSS on the reverse direction of each task would isolate this effect.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 3 minor

Summary. The paper proposes Target-Domain Structural Smoothing (TDSS), a plug-and-play module for unsupervised graph domain adaptation. TDSS applies a Laplacian smoothing loss on the target graph using neighborhoods obtained by k-hop or random-walk sampling, and combines it with the A2GNN classifier loss and an MMD alignment loss. The authors state a target-risk bound (Theorem 1, Eq. (7)) in which target risk is controlled by source risk, total variation distance, and model smoothness, and they claim that TDSS reduces target risk by improving model smoothness. Experiments on three ArnetMiner citation graphs across six transfer tasks report consistent improvements over existing UGDA baselines, including gains over A2GNN, and the code is publicly available.

Significance. If the empirical results hold, TDSS is an attractive baseline: it is simple, model-agnostic, and integrates into standard GNNs as well as specialized UGDA models. The paper includes a broad experimental study across six transfer tasks, three backbones, ablations, and hyperparameter sensitivity, and it releases code. However, the central theoretical claim is not currently established: the Laplacian smoothing loss is never shown to control the model-smoothness quantity that appears in the bound, and the bound itself contains an unexplained dependence on the smoothness parameter. The empirical superiority over A2GNN is plausible but cannot be verified from the reported numbers because no error bars, run counts, or significance-test details are given. The method's practical value may survive a corrected or reframed theory, but the manuscript's main justification requires substantial revision.

major comments (5)
  1. [Methodology, Eq. (4) vs. Definition 1 (Eq. (5)); Eq. (12) and Remark 1] The central theoretical claim is that TDSS reduces target risk by decreasing the model smoothness Φ, but no argument shows that LSR of Eq. (4) controls the Φ defined in Eq. (5). LSR is a degree-normalized quadratic penalty over edges of the sampled adjacency matrix eA of Eq. (3), while Φ is an expectation of a supremum over node pairs with graph distance at most k and feature distance at most r, evaluated on the model output f(θ,x). Three concrete gaps make the identification 2ΦT ↔ LSR unsupported: (i) eA retains only a subset of edges, so an unsampled edge can have a large output difference while LSR is zero; (ii) LSR is an edgewise penalty rather than a k-hop supremum, and no amplification argument from k=1 to general k is given; (iii) LSR is written for a feature representation f(x_i), not the classifier output f(θ,x), and no Lipschitz or other assumption links the two. Because Eq. (12) maps the bound's smoothness term to the proposed loss and Eq. (13) optimizes exactly those terms, the theory is post-hoc unless this connection is proved.
  2. [Remark 2 and Eq. (10)] The claim ΦS < ΦT is asserted without proof, and it is used to replace ΦS + ΦT with 2ΦT and to justify smoothing only the target graph. Even if the inequality held, it would not by itself show that the target-only regularizer controls the sum ΦS + ΦT; the source smoothness must also remain bounded during training. The paper provides no lemma, measurement, or experiment isolating this assumption. This is load-bearing because the operational loss in Eq. (13) smooths only the target graph.
  3. [Theorem 1, Eq. (9), and Appendix Eq. (20)] The covering-number term Z in the bound contains Φ^2 in the exponent, i.e., (2d)^{2Φ^2 Γ/r^2 + 1}. Covering numbers of the input space depend on dimension, radius, and diameter, not on the model's smoothness; no derivation of this Φ^2 dependence is provided. This makes K depend on Φ in an unexplained way and prevents the bound from being used as stated. The authors should either correct the expression or derive the exponent explicitly.
  4. [Methodology, Eqs. (2)-(3)] The random-walk sampling is claimed to explore multi-hop neighborhoods and avoid over-smoothing, but Eq. (3) sets eA_ij = 1 only if (v_i, v_j) ∈ E; any node visited by the walk without a direct edge is excluded. Thus the updated adjacency matrix contains only a subset of original edges, and the random walk reduces to edge subsampling rather than multi-hop neighborhood construction. This undermines the ablation contrast between 'w/o RW' (1-hop sampling) and the proposed random-walk variant, and it should be clarified or corrected.
  5. [Table 2] The asterisks claim significant improvements over A2GNN at the 0.05 level, but no standard deviations, number of independent runs, or test procedure are reported. With average gains of 2.04% Macro-F1 and 3.05% Micro-F1, the reader cannot verify that the differences are significant. The authors should report mean ± std over at least five seeds and specify the significance test used.
minor comments (3)
  1. [Table 3] In the A→D row, the GCN Original Micro-F1 entry is printed as '07055' and should read '0.7055'.
  2. [Introduction and Related Work] The contribution statement 'first to approach UGDA through model smoothness' is too strong given that SpecReg and A2GNN, both cited in the paper, already use smoothness-related or spectral regularization to tighten UGDA bounds; the claim should be softened.
  3. [Definition 2, Eq. (6)] The notation G1(v) and G2(v) for attribute distributions is used without a clear definition of what 'attribute distribution of node v' means, and the summation over v ∈ V in Eq. (6) is ambiguous because the two graphs may have different node sets; the definition should be made precise.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Theorem 1 is derived independently of the proposed loss; the asserted LSR–Φ link is an unsupported gap (correctness risk), not an equation-level reduction.

full rationale

The derivation chain in the appendix does not import LSR or Eq. (13) as assumptions. Theorem 1 follows from Wasserstein-robustness bounds (Yi et al.; Xu and Mannor), covering-number arguments, and TVD, with no step in which the target risk bound is defined in terms of LGC, LDA, or LSR. The underbrace alignment in Eq. (12) is post-hoc interpretation, not a definitional equivalence: Φ_T is an expected supremum over feature perturbations on the graph (Definition 1), whereas LSR is a degree-normalized quadratic penalty on the sampled adjacency matrix eA (Eq. 4). No equation equates the two, so the theory does not reduce to the method. Similarly, Remark 2's ΦS < ΦT is asserted, not derived, and the absence of a proposition bounding ΦT by LSR weakens the paper's theoretical justification, but an unsupported assertion is not circularity. The empirical comparisons in Tables 2 and 3 are against external baselines and do not depend on the theorem, so the central empirical claim is independently grounded. Minor self-citations (Chen et al. 2024a,b,c) appear only in general related-work context and are not load-bearing.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central theoretical claim rests on two ad hoc assumptions not derived in the paper: that the Laplacian smoothing loss reduces the feature-space smoothness Φ, and that ΦS < ΦT. The empirical method also depends on per-task hyperparameters α, β, λ, and γ, which are tuned on the evaluated transfer scenarios.

free parameters (5)
  • alpha (domain alignment weight) = 0.3 for several tasks; tuned in [0,1]
    Controls the strength of the MMD loss; sensitivity analysis shows task-specific optima.
  • beta (smoothing weight) = 0.2 for A->D, 0.3 for D->A; tuned in [0,1]
    Controls the strength of the Laplacian smoothing loss; optimal value varies by transfer task.
  • random walk length lambda = small, e.g., 2 for A->D
    Number of steps in each random walk; longer walks can cause over-smoothing and noise.
  • random walk count gamma = 3 to 5
    Number of random walks per node; moderate values balance information and redundancy.
  • k for k-hop sampling = not reported per task
    Alternative neighbor sampling strategy; k-hop variant underperforms random-walk variant.
assumptions (5)
  • domain assumption Source and target share label conditionals P(Ys|Xs) = P(Yt|Xt).
    Standard UGDA assumption stated in Problem Formulation; needed for classifier transfer.
  • ad hoc to paper The graph Laplacian smoothing loss LSR controls the feature-space smoothness Φ of Definition 1.
    Eq. (12) aligns LSR with Φ, but the appendix never derives this implication; LSR operates on graph adjacency while Φ is defined over feature-space perturbations.
  • ad hoc to paper The target domain requires more smoothing than the source domain, i.e., ΦS < ΦT.
    Remark 2 asserts this ordering to replace ΦS + ΦT with 2ΦT; no empirical or theoretical evidence is given.
  • standard math Covering number bound for the feature space under l-infinity norm.
    Used in the appendix to control the empirical process term; the expression N <= (2d)^{...} is taken from Vershynin but appears misstated.
  • standard math Wasserstein-ball distributional robustness is equivalent to l-infinity input robustness.
    Borrowed from Yi et al. 2021 as Eq. (17); the paper does not restate the theorem's conditions or verify its applicability to graph-structured data.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Smoothness Really Matters: A Simple Yet Effective Approach for Unsupervised Graph Domain Adaptation." pith.science (2026). https://pith.science/paper/W3QNSPFQ

@misc{pith2026241211654,
  author       = {Pith},
  title        = {Pith review of: Smoothness Really Matters: A Simple Yet Effective Approach for Unsupervised Graph Domain Adaptation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/W3QNSPFQ}},
  note         = {Machine review of arXiv:2412.11654}
}
read the original abstract

Unsupervised Graph Domain Adaptation (UGDA) seeks to bridge distribution shifts between domains by transferring knowledge from labeled source graphs to given unlabeled target graphs. Existing UGDA methods primarily focus on aligning features in the latent space learned by graph neural networks (GNNs) across domains, often overlooking structural shifts, resulting in limited effectiveness when addressing structurally complex transfer scenarios. Given the sensitivity of GNNs to local structural features, even slight discrepancies between source and target graphs could lead to significant shifts in node embeddings, thereby reducing the effectiveness of knowledge transfer. To address this issue, we introduce a novel approach for UGDA called Target-Domain Structural Smoothing (TDSS). TDSS is a simple and effective method designed to perform structural smoothing directly on the target graph, thereby mitigating structural distribution shifts and ensuring the consistency of node representations. Specifically, by integrating smoothing techniques with neighborhood sampling, TDSS maintains the structural coherence of the target graph while mitigating the risk of over-smoothing. Our theoretical analysis shows that TDSS effectively reduces target risk by improving model smoothness. Empirical results on three real-world datasets demonstrate that TDSS outperforms recent state-of-the-art baselines, achieving significant improvements across six transfer scenarios. The code is available in https://github.com/cwei01/TDSS.

Figures

Figures reproduced from arXiv: 2412.11654 by the authors.

Figure 1
Figure 1. An example of GNNs sensitivity to similar inputs. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Classification Micro-F1 comparisons between [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 4
Figure 4. Visualization of learned node embedding on [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figures from the paper (1 more)
Figure 3
Figure 3. Figure 3: The Micro-F1 performances of our TDSS w.r.t [PITH_FULL_IMAGE:figures/full_fig_p007_3.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 37 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Alon, U. 2007. Network motifs: theory and experimental approaches. Nature Reviews Genetics, 8(6): 450--461

  4. [4]

    Chen, P.; and Redner, S. 2010. Community structure of the physical review citation network. Journal of Informetrics, 4(3): 278--290

  5. [5]

    Chen, W.; Liu, Y.; Zhang, Z.; Zhuang, F.; and Zhong, J. 2024 a . Modeling Adaptive Inter-Task Feature Interactions via Sentiment-Aware Contrastive Learning for Joint Aspect-Sentiment Prediction. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 17781--17789

  6. [6]

    Chen, W.; Wu, Y.; Zhang, Z.; Zhuang, F.; He, Z.; Xie, R.; and Xia, F. 2024 b . FairGap: Fairness-aware Recommendation via Generating Counterfactual Graph. ACM Transactions on Information Systems, 42(4): 1--25

  7. [7]

    Chen, W.; Yuan, M.; Zhang, Z.; Xie, R.; Zhuang, F.; Wang, D.; and Liu, R. 2024 c . FairDgcl: Fairness-aware Recommendation with Dynamic Graph Contrastive Learning. arXiv preprint arXiv:2410.17555

  8. [8]

    Dai, Q.; Wu, X.-M.; Xiao, J.; Shen, X.; and Wang, D. 2022. Graph transfer learning via adversarial domain adaptation with graph convolution. IEEE Transactions on Knowledge and Data Engineering, 35(5): 4908--4922

Show all 57 references
  1. [9]

    Devroye, L.; Mehrabian, A.; and Reddad, T. 2018. The total variation distance between high-dimensional Gaussians with the same mean. arXiv preprint arXiv:1810.08693

  2. [10]

    Field, D. A. 1988. Laplacian smoothing and Delaunay triangulations. Communications in applied numerical methods, 4(6): 709--712

  3. [11]

    Filippone, M.; Camastra, F.; Masulli, F.; and Rovetta, S. 2008. A survey of kernel and spectral methods for clustering. Pattern recognition, 41(1): 176--190

  4. [12]

    Gao, C.; Zheng, Y.; Li, N.; Li, Y.; Qin, Y.; Piao, J.; Quan, Y.; Chang, J.; Jin, D.; He, X.; et al. 2023. A survey of graph neural networks for recommender systems: Challenges, methods, and directions. ACM Transactions on Recommender Systems, 1(1): 1--51

  5. [13]

    Grover, A.; and Leskovec, J. 2016. node2vec: Scalable feature learning for networks. In Proceedings of the 22nd ACM SIGKDD international conference on Knowledge discovery and data mining, 855--864

  6. [14]

    Hamilton, W.; Ying, Z.; and Leskovec, J. 2017. Inductive representation learning on large graphs. Advances in neural information processing systems, 30

  7. [15]

    E.; and Roweis, S

    Hinton, G. E.; and Roweis, S. 2002. Stochastic neighbor embedding. Advances in neural information processing systems, 15

  8. [16]

    N.; and Welling, M

    Kipf, T. N.; and Welling, M. 2016. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907

  9. [17]

    Li, H.; Wang, X.; Zhang, Z.; and Zhu, W. 2022. Ood-gnn: Out-of-distribution generalized graph neural network. IEEE Transactions on Knowledge and Data Engineering, 35(7): 7328--7340

  10. [18]

    Li, J.; Yu, Z.; Du, Z.; Zhu, L.; and Shen, H. T. 2024. A comprehensive survey on source-free domain adaptation. IEEE Transactions on Pattern Analysis and Machine Intelligence

  11. [19]

    Li, W.; and Chen, S. 2022. Partial domain adaptation without domain alignment. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(7): 8787--8797

  12. [20]

    Liang, K.; Meng, L.; Liu, M.; Liu, Y.; Tu, W.; Wang, S.; Zhou, S.; Liu, X.; Sun, F.; and He, K. 2024 a . A survey of knowledge graph reasoning on graph types: Static, dynamic, and multi-modal. IEEE Transactions on Pattern Analysis and Machine Intelligence

  13. [21]

    Liang, K.; Meng, L.; Zhou, S.; Tu, W.; Wang, S.; Liu, Y.; Liu, M.; Zhao, L.; Dong, X.; and Liu, X. 2024 b . MINES: Message Intercommunication for Inductive Relation Reasoning over Neighbor-Enhanced Subgraphs. In Proceedings of the AAAI Conference on Artificial Intelligence, vo...

  14. [22]

    Liao, W.; Zhu, Y.; Li, Y.; Zhang, Q.; Ou, Z.; and Li, X. 2024. RevGNN: Negative Sampling Enhanced Contrastive Graph Learning for Academic Reviewer Recommendation. ACM Transactions on Information Systems, 43(1): 1--26

  15. [23]

    Liu, M.; Fang, Z.; Zhang, Z.; Gu, M.; Zhou, S.; Wang, X.; and Bu, J. 2024 a . Rethinking propagation for unsupervised graph domain adaptation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 13963--13971

  16. [24]

    Liu, S.; Li, T.; Feng, Y.; Tran, N.; Zhao, H.; Qiu, Q.; and Li, P. 2023. Structural re-weighting improves graph domain adaptation. In International Conference on Machine Learning, 21778--21793. PMLR

  17. [25]

    Liu, S.; Zou, D.; Zhao, H.; and Li, P. 2024 b . Pairwise Alignment Improves Graph Domain Adaptation. arXiv preprint arXiv:2403.01092

  18. [26]

    Luo, H.; Zhuang, F.; Xie, R.; Zhu, H.; Wang, D.; An, Z.; and Xu, Y. 2024. A survey on causal inference for recommendation. The Innovation, 5(2): 100590

  19. [27]

    Nikolentzos, G.; and Vazirgiannis, M. 2020. Random walk graph neural networks. Advances in Neural Information Processing Systems, 33: 16211--16222

  20. [28]

    Perozzi, B.; Al-Rfou, R.; and Skiena, S. 2014. Deepwalk: Online learning of social representations. In Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining, 701--710

  21. [29]

    Qiao, Z.; Luo, X.; Xiao, M.; Dong, H.; Zhou, Y.; and Xiong, H. 2023. Semi-supervised domain adaptation in graph transfer learning. In Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence, 2279--2287

  22. [30]

    I Can't Believe It's Not Better!

    Rosca, M.; Weber, T.; Gretton, A.; and Mohamed, S. 2020. A case for new neural network smoothness constraints. In Zosa Forde, J.; Ruiz, F.; Pradier, M. F.; and Schein, A., eds., Proceedings on "I Can't Believe It's Not Better!" at NeurIPS Workshops, volume 137 of Proceedings o...

  23. [31]

    R \"u schendorf, L. 1985. The Wasserstein distance and approximation theorems. Probability Theory and Related Fields, 70(1): 117--129

  24. [32]

    Shen, X.; Dai, Q.; Chung, F.-l.; Lu, W.; and Choi, K.-S. 2020 a . Adversarial deep network embedding for cross-network node classification. In Proceedings of the AAAI conference on artificial intelligence, volume 34, 2991--2999

  25. [33]

    Shen, X.; Dai, Q.; Mao, S.; Chung, F.-l.; and Choi, K.-S. 2020 b . Network together: Node classification via cross-network deep network embedding. IEEE Transactions on Neural Networks and Learning Systems, 32(5): 1935--1948

  26. [34]

    Shi, B.; Wang, Y.; Guo, F.; Xu, B.; Shen, H.; and Cheng, X. 2024. Graph Domain Adaptation: Challenges, Progress and Prospects. arXiv preprint arXiv:2402.00904

  27. [35]

    Tang, J.; Zhang, J.; Yao, L.; Li, J.; Zhang, L.; and Su, Z. 2008. Arnetminer: extraction and mining of academic social networks. In Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining, 990--998

  28. [36]

    W.; Wellner, J

    Van Der Vaart, A. W.; Wellner, J. A.; van der Vaart, A. W.; and Wellner, J. A. 1996. Weak convergence. Springer

  29. [37]

    Velickovic, P.; Cucurull, G.; Casanova, A.; Romero, A.; Lio, P.; Bengio, Y.; et al. 2017. Graph attention networks. stat, 1050(20): 10--48550

  30. [38]

    Vershynin, R. 2018. High-dimensional probability: An introduction with applications in data science, volume 47. Cambridge university press

  31. [39]

    Villani, C.; et al. 2009. Optimal transport: old and new, volume 338. Springer

  32. [40]

    Wang, R.; and Manchester, I. 2023. Direct parameterization of lipschitz-bounded deep networks. In International Conference on Machine Learning, 36093--36110. PMLR

  33. [41]

    Wang, Y.; Wang, D.; Liu, H.; Hu, B.; Yan, Y.; Zhang, Q.; and Zhang, Z. 2024. Optimizing Long-tailed Link Prediction in Graph Neural Networks through Structure Representation Enhancement. arXiv preprint arXiv:2407.20499

  34. [42]

    Wilson, G.; and Cook, D. J. 2020. A survey of unsupervised deep domain adaptation. ACM Transactions on Intelligent Systems and Technology (TIST), 11(5): 1--46

  35. [43]

    Wu, F.; Souza, A.; Zhang, T.; Fifty, C.; Yu, T.; and Weinberger, K. 2019. Simplifying graph convolutional networks. In International conference on machine learning, 6861--6871. PMLR

  36. [44]

    Wu, J.; He, J.; and Ainsworth, E. 2023. Non-iid transfer learning on graphs. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 10342--10350

  37. [45]

    Wu, M.; Pan, S.; Zhou, C.; Chang, X.; and Zhu, X. 2020. Unsupervised domain adaptive graph convolutional networks. In Proceedings of the web conference 2020, 1457--1467

  38. [46]

    Xiao, Z.; Wang, H.; Jin, Y.; Feng, L.; Chen, G.; Huang, F.; and Zhao, J. 2023. SPA: a graph spectral alignment perspective for domain adaptation. Advances in Neural Information Processing Systems, 36

  39. [47]

    Xu, B.; Zhao, C.; Jiang, W.; Zhu, P.; Dai, S.; Pang, C.; Sun, Z.; Wang, S.; and Sun, Y. 2023. Retrieval-augmented domain adaptation of language models. In Proceedings of the 8th Workshop on Representation Learning for NLP (RepL4NLP 2023), 54--64

  40. [48]

    Xu, H.; and Mannor, S. 2012. Robustness and generalization. Machine learning, 86: 391--423

  41. [49]

    Xu, K.; Ma, Z.; Xu, L.; He, G.; Li, Y.; Yu, W.; Han, T.; and Yang, C. 2024 a . An End-to-End Real-World Camera Imaging Pipeline. In ACM Multimedia 2024

  42. [50]

    Xu, K.; Xu, L.; He, G.; Yu, W.; and Li, Y. 2024 b . Beyond Alignment: Blind Video Face Restoration via Parsing-Guided Temporal-Coherent Transformer. In Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, \ IJCAI-24 \ , 1489--1497

  43. [51]

    Yi, M.; Hou, L.; Sun, J.; Shang, L.; Jiang, X.; Liu, Q.; and Ma, Z. 2021. Improved ood generalization via adversarial training and pretraing. In International Conference on Machine Learning, 11987--11997. PMLR

  44. [52]

    You, Y.; Chen, T.; Wang, Z.; and Shen, Y. 2023. Graph domain adaptation via theory-grounded spectral regularization. In The eleventh international conference on learning representations

  45. [53]

    Zhang, P. 2015. Evaluating accuracy of community detection using the relative normalized mutual information. Journal of Statistical Mechanics: Theory and Experiment, 2015(11): P11006

  46. [54]

    Zhang, Z.; Yang, H.; Bu, J.; Zhou, S.; Yu, P.; Zhang, J.; Ester, M.; and Wang, C. 2018. ANRL: attributed network representation learning via deep neural networks. In Ijcai, volume 18, 3155--3161

  47. [55]

    Zhu, H.; Xiong, F.; Chen, H.; Xiong, X.; and Wang, L. 2024. Incorporating a triple graph neural network with multiple implicit feedback for social recommendation. ACM Transactions on the Web, 18(2): 1--26

  48. [56]

    Zhu, Y.; Cong, F.; Zhang, D.; Gong, W.; Lin, Q.; Feng, W.; Dong, Y.; and Tang, J. 2023. Wingnn: Dynamic graph neural networks with random gradient aggregation window. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 3650--3662

  49. [57]

    Zhuang, F.; Qi, Z.; Duan, K.; Xi, D.; Zhu, Y.; Zhu, H.; Xiong, H.; and He, Q. 2020. A comprehensive survey on transfer learning. Proceedings of the IEEE, 109(1): 43--76

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.