Pith. sign in

REVIEW 5 major objections 6 minor 50 references

A Unified Framework for Interactive Visual Graph Matching via Attribute-Structure Synchronization

T0 review · 5 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read Graph matching improves when structure and attributes are synchronized through canonical correlation analysis rather than concatenated.

desk verdict Plausible visual graph matching system, but the quantitative evaluation does not support the headline superiority claim. read the letter →

arxiv 2507.19750 v1 pith:3JBDTSG2 submitted 2025-07-26 cs.IR

classification cs.IR
keywords graphmatchingcanonicalcorrelationanalysisrepresentationlearningvisualanalyticsattribute-structuresynchronizationk-nearestneighborqueryinteractivevisualization
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

Graph retrieval traditionally compares structures, but real graphs carry node attributes that also matter for similarity. This paper claims that when both are projected into one shared embedding space by canonical correlation analysis (CCA), the resulting k-nearest-neighbor matches preserve structure and attributes better than using either alone or simply concatenating the two feature vectors. The paper wraps this synchronization in an interactive visual system that lets users choose a target graph from a projection, sketch a structure, or set attribute ranges, and then inspect matches through node-link diagrams and parallel coordinates. If the claim holds, graph databases gain a modular and interpretable way to query by structure and semantics at once, without writing graph query languages.

What carries the argument

The load-bearing mechanism is CCA-based attribute-structure synchronization: canonical correlation analysis finds linear projection directions for the structure feature space and the attribute feature space such that the projected variables are maximally correlated, and the two projected vectors are concatenated into a unified fused embedding. The same mechanism performs joint dimensionality reduction, so the matching space is low-dimensional and distance-based. Graph2vec supplies the structural side of the pairing by embedding each whole graph from rooted subgraphs generated through Weisfeiler-Lehman relabeling, while statistical aggregation of node attributes supplies the attribute side.

What would settle it

Build a synthetic graph collection in which the matching relation between structure and attributes is deliberately non-linear (for example, graphs match only when the structure class matches the parity of an attribute value), run the same k-nearest-neighbor matching, and check whether the CCA fused space systematically ranks mismatched graphs closer than a simple non-linear alignment does.

Watch

Extended reading notes

Core claim

The central discovery is that attribute-structure synchronization via CCA yields credible graph matching results. Given a set of graphs, the paper first embeds each graph's topology with Graph2vec to obtain a structure vector and builds an attribute vector from node statistics, then applies CCA to these paired vectors so that the projections of structure and attributes are maximally correlated. The two projected vectors are concatenated into a fused representation, and graphs are matched by k-nearest neighbors in that fused space. In quantitative comparisons on a genealogy dataset and a co-author network dataset, this method reports lower average graph edit distance and lower average attribute Euclidean distance among matches than direct concatenation and indirect concatenation baselines, and it balances the two similarity types better than structure-only or attribute-only matching. The paper also demonstrates the framework's use through two case studies in which experts retrieve family trees and research collaboration communities.

Load-bearing premise

The load-bearing premise is that the relationship between the Graph2vec structure embeddings and the hand-selected attribute vectors is sufficiently linear that CCA's projection places genuinely matching graphs near each other; the paper acknowledges in its discussion that non-linear or high-order dependencies are not modeled.

Editorial extensions

If this is right

  • Retrieval systems can fold node attributes into graph similarity without designing a custom fused metric, because CCA learns the projection from the data itself.
  • On the two real-world datasets tested, the fused space outperforms direct and indirect concatenation on both structure and attribute similarity, so how attributes are fused matters, not just that they are included.
  • Because matching is distance-based in a low-dimensional space, the approach remains fast enough for interactive exploration, with running times comparable to the baselines.
  • Non-expert users can specify a target graph by selecting a point in the projection, drawing a structure, or adjusting attribute sliders, replacing graph query language with visual query definition.
  • The modular separation of structure encoding and cross-view alignment means either component can be swapped, so the framework can adopt newer encoders or alignment models without redesign.

Reading between the lines

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

  • A direct test of the modular claim would swap Graph2vec for a graph neural network encoder and check whether the CCA advantage over concatenation survives; the paper's reasoning implies it should, since the alignment mechanism, not the encoder, carries the benefit.
  • The paper's reported trade-off suggests that on datasets where structure and attributes are weakly correlated, the CCA fused space should drift toward the behavior of the weaker view; measuring match quality across a range of synthetic structure-attribute correlations would make that boundary explicit.
  • A task-based user study could quantify whether the evaluation views actually improve decision speed or accuracy; the paper claims interpretability but does not measure it formally.
  • The unweighted concatenation of the two CCA projections gives structure and attributes equal say; introducing a user-controlled weighting, mentioned as future work, would let the framework tune the trade-off without retraining the alignment.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper proposes a visual analytics framework for graph-level matching that combines structural embeddings from Graph2vec with manually constructed attribute vectors through Canonical Correlation Analysis (CCA), then uses distance-based k-NN matching in the fused space. The system includes interactive views for query specification, projection, and evaluation. The authors report quantitative comparisons against structure-only, attribute-only, direct-concatenation, and indirect-concatenation baselines on two real-world datasets, together with case studies and expert feedback.

Significance. If the quantitative claims were robust, the CCA-based fusion would offer a useful modular and interpretable alternative to end-to-end deep fusion for interactive graph retrieval: it decouples structure embedding from attribute alignment, uses external evaluation metrics (GED and attribute Euclidean distance) that are not fitted parameters, and the visual system addresses a real usability gap in graph querying. The paper is honest about the linearity limitation of CCA and about the lack of formal user evaluation. However, the central empirical claim of superiority over baselines is not supported by the evidence as presented, which limits the paper's current significance.

major comments (5)
  1. [§6.2, Table 1] The headline claim that the proposed method 'almost performs better than DC and IDC' is not supported by the reported numbers. In Genealogy k=20, Str-Sim is 10.01 for Our versus 9.64 for Str and 9.74 for DC, so the method is worse than both on structure similarity; in every Genealogy row, Attr-Sim for Our (12.53, 13.04, 15.36) is worse than Attr (7.02, 7.97, 10.20). No error bars, standard deviations, or significance tests are reported, and the reader cannot tell whether differences of 0.1–0.3 are noise. The claim 'proves that our method effectively overcomes the shortcomings of conventional fusion methods' should be replaced by a weaker claim or supported by paired statistical testing across multiple random target selections and Graph2vec training runs.
  2. [§6.2, Table 1] The IDC value 40.06 for Genealogy k=20 Str-Sim is an order of magnitude larger than all other entries and is not explained anywhere. If this reflects a scaling or normalization artifact in the IDC implementation, that baseline should be fixed or excluded; as presented, the comparison with IDC is not meaningful, and the later statement that Our is better than IDC relies on this anomalous cell.
  3. [§4.2 and §7.1] The fusion model assumes a linear relationship between Graph2vec structure embeddings and hand-chosen attribute vectors, and §7.1 explicitly acknowledges that CCA cannot model non-linear dependencies. The paper does not test whether this linearity assumption holds on the two datasets, and the only nonlinear alternative, KCCA, is dismissed in the text with results relegated to a supplementary file that is not provided. Without either a linearity diagnostic or a reproducible nonlinear baseline comparison, the generalizability claim for the synchronization method remains unsubstantiated.
  4. [§6.2, Table 2] Table 2 is presented as a robustness evaluation, but it repeats the same k values (5, 10, 15) already shown in Table 1 and adds only point estimates for KCCA with no variance or significance information. The claim of 'consistently' low error across k cannot be assessed from single-run point estimates, and no new variation (e.g., different random seeds, different target subsets, or different Graph2vec hyperparameters) is actually varied.
  5. [§6.3 and §7.2] The interactive system is a core contribution, but its effectiveness is supported only by qualitative expert anecdotes. Section 7.2 concedes that the visual interface's effectiveness 'has not been formally quantified.' Since the paper's stated contributions include 'user-friendly interactions' and 'convenience,' a task-based user study or at least a structured log analysis is needed to support these claims; otherwise they should be presented as design proposals rather than validated results.
minor comments (6)
  1. [§7 title] The section title is misspelled as 'Disccusion'; it should be 'Discussion.'
  2. [§4.2, Eq. (4)] The sentence preceding Eq. (4) refers to 'The first pair of these basis vectors, {hAi}' but should be '{hS1, hA1}'; this looks like a typographical error.
  3. [§6.3] There are several typographical and wording issues, such as 'single-linage' for 'single-lineage,' 'traditionary' for 'traditional,' and the quotation containing 'He clicked... and found that Liu's co-author networks...' where the quote appears to mix narration with direct speech.
  4. [Abstract] The abstract contains a leftover LaTeX macro '\revise{our method}' that should be resolved to plain text.
  5. [References] Reference [34] is garbled: the author list contains 'Francis, nadime and green, alastair and guagliardo, paolo and libkin, leonid...' and needs to be corrected to the actual authors of the cited Cypher paper.
  6. [§6.2] The definitions of DC and IDC are underspecified: it is not stated which dimensionality reduction method is used for the joint reduction or for the 'unified dimension' step, which matters because the anomalous IDC value may stem from this choice.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the CCA attribute-structure synchronization is evaluated by external GED and attribute-distance metrics, and no prediction reduces to a fitted parameter or to a load-bearing self-citation.

full rationale

The paper's derivation chain is not circular. Structure vectors come from Graph2vec and attribute vectors are handcrafted; CCA then solves for projection matrices H_S and H_A by maximizing canonical correlation (Eq. 4). The resulting fused vectors are used for k-NN retrieval, and the reported quality metrics are average graph edit distance (GED, ref. [49]) and average Euclidean distance between raw attribute vectors (ref. [50]). Neither evaluation metric appears in the CCA objective, in the projection construction, or in the fusion objective, so the proposed method's reported Str-Sim and Attr-Sim values are not forced by the fitting procedure. The comparison against Str, Attr, DC, and IDC is an external empirical comparison, not a renaming of the method's own training criterion. The paper's self-references, such as [26]-[28], are contextual related-work citations about graph representation learning and do not supply the uniqueness of the CCA construction or the matching result. The admitted limitation that CCA cannot model nonlinear dependencies (Sec. 7.1) is a scope statement, not a circular justification. The absence of error bars, single-run point estimates, and in-sample evaluation are important statistical validity concerns, but they are not instances of the derivation reducing to its inputs; therefore they do not raise the circularity score.

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

The framework introduces no new physical or conceptual entities. Its free parameters are the unspecified dimensions and hyperparameters of Graph2vec and CCA, plus the hand-chosen attribute sets and neighborhood size. The axioms are mostly domain assumptions about the adequacy of Graph2vec embeddings, manual attributes, and linear CCA for capturing matching-relevant structure and semantics. These are standard engineering choices for a modular visual analytics pipeline, but they are not derived from first principles or independently validated.

free parameters (5)
  • CCA projection dimension m = not reported
    The number of canonical components retained for the fused embedding is never stated; it directly controls the dimensionality of the matching space and is a free modeling choice.
  • Graph2vec embedding dimension N_s = not reported
    The structural feature dimension is a hyperparameter of Graph2vec that is not specified in the paper, though it determines the CCA input.
  • Graph2vec training hyperparameters = not reported
    Context size, epochs, and learning rate are not given, so a re-implementation cannot match the reported embeddings.
  • Hand-selected attribute sets per dataset = genealogy: TS, AA, PN, VN, AG; co-author: AU, CI, PA, YE, RA, WE
    The attribute vectors are chosen manually by the authors and domain experts; different attribute choices would change matching behavior, and these are not derived from a principled feature-selection procedure.
  • k in k-NN matching = 20/40/60 (genealogy), 5/10/15 (co-author)
    The number of nearest neighbors is user-specified and varies per dataset; the reported comparisons depend on these values.
assumptions (5)
  • domain assumption Graph2vec embeddings faithfully represent global graph structure relevant to matching
    The method relies on Graph2vec's graph-level vectors as the sole structural representation; the paper itself notes in Section 7.1 that it may not capture finer-grained topological semantics.
  • domain assumption The hand-selected attribute vectors capture the semantic information users care about for graph matching
    Five (genealogy) and six (co-author) manually defined attributes are assumed sufficient; Section 4.2 states attributes are extracted based on domain knowledge.
  • domain assumption Linear canonical correlations between structure and attribute spaces are a meaningful basis for graph similarity
    CCA assumes linearity and that maximizing canonical correlation places similar graphs close together; Section 7.1 admits limited capacity for non-linear dependencies.
  • domain assumption Nearest neighbors in the CCA-fused space correspond to graphs that balance structural and attribute similarity
    The matching scheme uses k-NN or cluster membership in the fused space; this equivalence is assumed, not proven, and is the core of the evaluation.
  • standard math Eigenvalue problem for CCA is solved correctly with standard numerical methods
    Section 4.2 describes the standard CCA eigenvalue formulation; this is a textbook result.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Unified Framework for Interactive Visual Graph Matching via Attribute-Structure Synchronization." pith.science (2026). https://pith.science/paper/3JBDTSG2

@misc{pith2026250719750,
  author       = {Pith},
  title        = {Pith review of: A Unified Framework for Interactive Visual Graph Matching via Attribute-Structure Synchronization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3JBDTSG2}},
  note         = {Machine review of arXiv:2507.19750}
}
read the original abstract

In traditional graph retrieval tools, graph matching is commonly used to retrieve desired graphs from extensive graph datasets according to their structural similarities. However, in real applications, graph nodes have numerous attributes which also contain valuable information for evaluating similarities between graphs. Thus, to achieve superior graph matching results, it is crucial for graph retrieval tools to make full use of the attribute information in addition to structural information. We propose a novel framework for interactive visual graph matching. In the proposed framework, an attribute-structure synchronization method is developed for representing structural and attribute features in a unified embedding space based on Canonical Correlation Analysis (CCA). To support fast and interactive matching, \revise{our method} provides users with intuitive visual query interfaces for traversing, filtering and searching for the target graph in the embedding space conveniently. With the designed interfaces, the users can also specify a new target graph with desired structural and semantic features. Besides, evaluation views are designed for easy validation and interpretation of the matching results. Case studies and quantitative comparisons on real-world datasets have demonstrated the superiorities of our proposed framework in graph matching and large graph exploration.

Figures

Figures reproduced from arXiv: 2507.19750 by the authors.

Figure 1
Figure 1. The pipeline of our graph matching system based on attribute￾structure synchronization. Users first import a dataset and configure the matching model via the control panel. A target graph can be selected from the projection view or constructed manually. The selected graph is dis￾played in the target view for detailed inspection. Once matching is trig￾gered, candidate graphs with similar structure and attributes are … view at source ↗
Figure 2
Figure 2. Given a set of n graphs (a), we extract the topological feature vec [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. A case study is conducted to retrieve research communities with similar collaboration and attributes from a large amount of co-author network [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Exploration on the projection view (a) with structure features and [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Given the target graph in Figure 6(b), the matching results under di [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Two graphs with special features are selected from the 2D attribute [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Definition and matching of a target graph. (a) and (b) enable users [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Comparison of four matching results in Figure 7(c). [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 50 canonical work pages

  1. [1]

    D. Liu, G. Yang, Y . Wang, H. Jin, E. Chen, How to protect ourselves from overlapping community detection in social networks, IEEE Transactions on Big Data 8 (4) (2022) 894–904

  2. [2]

    Enterprise Analytics using Graph Database and Graph-based Deep Learning

    S. Henna, S. K. Kalliadan, Enterprise analytics using graph database and graph-based deep learning, arXiv preprint arXiv:2108.02867 (2021)

  3. [3]

    S. Ji, S. Pan, E. Cambria, P. Marttinen, S. Y . Philip, A survey on knowl- edge graphs: Representation, acquisition, and applications, IEEE transac- tions on neural networks and learning systems 33 (2) (2021) 494–514

  4. [4]

    N. M. Kriege, F. D. Johansson, C. Morris, A survey on graph kernels, Applied Network Science 5 (1) (2020) 1–42

  5. [5]

    J. Song, J. Song, X. Yuan, X. He, X. Zhu, Graph representation-based deep multi-view semantic similarity learning model for recommendation, Future Internet 14 (2) (2022) 32

  6. [6]

    Z. Lan, B. Hong, Y . Ma, F. Ma, More interpretable graph similarity com- putation via maximum common subgraph inference, IEEE Transactions on Knowledge and Data Engineering 36 (11) (2024) 6588–6599

  7. [7]

    Saadah, M

    S. Saadah, M. F. E. Saputro, K. R. S. Wiharja, Graph edit distance op- timized with greedy algorithm for similarity search in business process graphs, in: 2020 International Conference on Data Science and Its Appli- cations (ICoDSA), 2020, pp. 1–7

  8. [8]

    Mahmood, K

    S. Mahmood, K. Mueller, Interactive subspace cluster analysis guided by semantic attribute associations, IEEE Transactions on Visualization and Computer Graphics (2023)

Show all 50 references
  1. [9]

    P. Zhou, Y . Zhang, F. Chen, K. Pang, M. Lu, Heterogeneous ppi net- work representation learning for protein complex identification, in: In- ternational Symposium on Bioinformatics Research and Applications, Springer, 2022, pp. 217–228

  2. [10]

    Z. Deng, D. Weng, J. Chen, R. Liu, Z. Wang, J. Bao, Y . Zheng, Y . Wu, Airvis: Visual analytics of air pollution propagation, IEEE transactions on visualization and computer graphics 26 (1) (2019) 800–810

  3. [11]

    Wangmo, L

    C. Wangmo, L. Wiese, Subtempora: A hybrid approach for optimising subgraph searching, in: International Conference on Data Management Technologies and Applications, Springer, 2021, pp. 66–89. 12/Computers & Graphics (2025)

  4. [12]

    F. Bi, L. Chang, X. Lin, L. Qin, W. Zhang, Efficient subgraph matching by postponing cartesian products, in: Proceedings of the 2016 International Conference on Management of Data, 2016, pp. 1199–1214

  5. [13]

    C. Q. Cheng, K. S. Wong, L. K. Soon, l2match: Optimization techniques on subgraph matching algorithm using label pair, neighboring label index, and jump-redo method, in: 2024 International Conference on Electronics, Information, and Communication (ICEIC), IEEE, 2024, pp. 1–6

  6. [14]

    Karpov, Q

    N. Karpov, Q. Zhang, Syncsignature: a simple, efficient, parallelizable framework for tree similarity joins, Proceedings of the VLDB Endow- ment 16 (2) (2022) 330–342

  7. [15]

    Y . Chai, J. Li, Q. Zhang, J. Ge, X. Wang, Asm: Adaptive subgraph match- ing via efficient compression and label filter, in: Asia-Pacific Web (AP- Web) and Web-Age Information Management (W AIM) Joint International Conference on Web and Big Data, Springer, 2024, pp. 30–42

  8. [16]

    Y . Wang, C. Jin, S. Cai, Pathlad+: Towards effective exact methods for subgraph isomorphism problem, Artificial Intelligence 337 (2024) 104219

  9. [17]

    C. Q. Cheng, K. S. Wong, L. K. Soon, l2match: Optimization techniques on subgraph matching algorithm using label pair, neighboring label index, and jump-redo method, arXiv preprint arXiv:2311.16603 (2023)

  10. [18]

    W. Yang, C. Chen, J. Zhu, L. Li, P. Liu, S. Liu, A survey of visual analytics research for improving training data quality, Journal of Computer-Aided Design & Computer Graphics 35 (11) (2023) 1629–1642

  11. [19]

    F. Lyu, C. Chen, J. Zhang, X. Feng, Z. Tang, Visualization for supercom- puter system: A survey, Journal of Computer-Aided Design & Computer Graphics 36 (3) (2024) 321–335

  12. [20]

    Pienta, A

    R. Pienta, A. Tamersoy, A. Endert, S. Navathe, H. Tong, D. H. Chau, Visage: Interactive visual graph querying, in: Proceedings of the Interna- tional Working Conference on Advanced Visual Interfaces, 2016

  13. [21]

    Cao, Y .-R

    N. Cao, Y .-R. Lin, L. Li, H. Tong, g-miner: Interactive visual group min- ing on multivariate graphs, in: Proceedings of the 33rd Annual ACM Con- ference on Human Factors in Computing Systems, 2015

  14. [22]

    P. C. Wong, D. Haglin, D. Gillen, D. Chavarria, V . Castellana, C. Joslyn, A. Chappell, S. Zhang, A visual analytics paradigm enabling trillion-edge graph exploration, in: 2015 IEEE 5th Symposium on Large Data Analysis and Visualization (LDA V), IEEE, 2015

  15. [23]

    W. Chen, F. Guo, D. Han, J. Pan, X. Nie, J. Xia, X. Zhang, Structure- based suggestive exploration: a new approach for effective exploration of large networks, IEEE transactions on visualization and computer graphics 25 (1) (2018) 555–565

  16. [24]

    Pienta, F

    R. Pienta, F. Hohman, A. Endert, A. Tamersoy, K. Roundy, C. Gates, S. Navathe, D. H. Chau, Vigor: interactive visual exploration of graph query results, IEEE transactions on visualization and computer graphics 24 (1) (2017) 215–225

  17. [25]

    H. Song, Z. Dai, P. Xu, L. Ren, Interactive visual pattern search on graph data via graph representation learning, IEEE Transactions on Visualiza- tion and Computer Graphics 28 (1) (2021) 335–345

  18. [26]

    Z. Zhou, C. Shi, X. Shen, L. Cai, H. Wang, Y . Liu, Y . Zhao, W. Chen, Context-aware sampling of large networks via graph representation learn- ing, IEEE Transactions on Visualization and Computer Graphics 27 (2) (2020) 1709–1719

  19. [27]

    Z. Zhou, H. Wang, Z. Zhao, F. Zheng, Y . Wang, W. Chen, Y . Wang, Chartkg: A knowledge-graph-based representation for chart images, IEEE Transactions on Visualization and Computer Graphics (2024)

  20. [28]

    Z. Zhou, L. Sun, H. Wang, W. Yu, Y . Liu, X. Zhang, Y . Wang, W. Chen, imgc: Interactive multiple graph clustering with constrained laplacian rank, IEEE Transactions on Human-Machine Systems 53 (2) (2022) 427– 437

  21. [29]

    M. Grohe, word2vec, node2vec, graph2vec, x2vec: Towards a theory of vector embeddings of structured data, in: Proceedings of the 39th ACM SIGMOD-SIGACT-SIGAI Symposium on Principles of Database Sys- tems, 2020, pp. 1–16

  22. [30]

    M. Tang, C. Yang, P. Li, Graph auto-encoder via neighborhood wasser- stein reconstruction, arXiv preprint arXiv:2202.09025 (2022)

  23. [31]

    R. Yang, J. Shi, X. Xiao, Y . Yang, S. S. Bhowmick, J. Liu, Pane: scalable and effective attributed network embedding, The VLDB Journal 32 (6) (2023) 1237–1262

  24. [32]

    J. Wang, J. Guo, Y . Sun, J. Gao, S. Wang, Y . Yang, B. Yin, Dgnn: Decou- pled graph neural networks with structural consistency between attribute and graph embedding representations, IEEE Transactions on Big Data (2024)

  25. [33]

    C. Wang, J. Han, Dl4scivis: A state-of-the-art survey on deep learning for scientific visualization, IEEE Transactions on Visualization and Com- puter Graphics (2022)

  26. [34]

    H. Li, H. Wang, Z. Yang, M. Odagaki, Francis, nadime and green, alas- tair and guagliardo, paolo and libkin, leonid and lindaaker, tobias and marsault, victor and plantikow, stefan and rydberg, mats and selmer, petra and taylor, andr ´es, in: Proceedings of the 2018 internatio...

  27. [35]

    Thakkar, D

    H. Thakkar, D. Punjani, S. Auer, M.-E. Vidal, Towards an integrated graph algebra for graph pattern matching with gremlin, in: D. Bensli- mane, E. Damiani, W. I. Grosky, A. Hameurlain, A. Sheth, R. R. Wagner (Eds.), Database and Expert Systems Applications, Springer Internatio...

  28. [36]

    Hogan, A

    A. Hogan, A. Hogan, Sparql query language, The Web of Data (2020) 323–448

  29. [37]

    A. Khan, Y . Wu, C. C. Aggarwal, X. Yan, Nema: Fast graph search with label similarity, Proceedings of the VLDB Endowment 6 (3) (2013) 181– 192

  30. [38]

    P. Zhao, J. Han, On graph query optimization in large networks, Proceed- ings of the VLDB Endowment 3 (1-2) (2010) 340–351

  31. [39]

    K. D. Doan, S. Manchanda, S. Mahapatra, C. K. Reddy, Interpretable graph similarity computation via differentiable optimal alignment of node embeddings, in: Proceedings of the 44th international ACM SIGIR con- ference on research and development in information retrieval, 2021,...

  32. [40]

    O.-H. Kwon, T. Crnovrsanin, K.-L. Ma, What would a graph look like in this layout? a machine learning approach to large graph visualization, IEEE transactions on visualization and computer graphics 24 (1) (2017) 478–488

  33. [41]

    H. Li, Y . Wang, A. Wu, H. Wei, H. Qu, Structure-aware visualization retrieval, in: Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems, 2022

  34. [42]

    M. E. Sargin, Y . Yemez, E. Erzin, A. M. Tekalp, Audiovisual synchroniza- tion and fusion using canonical correlation analysis, IEEE transactions on Multimedia 9 (7) (2007) 1396–1403

  35. [43]

    Shervashidze, P

    N. Shervashidze, P. Schweitzer, E. J. Van Leeuwen, K. Mehlhorn, K. M. Borgwardt, Weisfeiler-lehman graph kernels., Journal of Machine Learn- ing Research 12 (9) (2011)

  36. [44]

    Ester, H.-P

    M. Ester, H.-P. Kriegel, J. Sander, X. Xu, et al., A density-based algorithm for discovering clusters in large spatial databases with noise, in: kdd, V ol. 96, 1996

  37. [45]

    J. MacQueen, et al., Some methods for classification and analysis of mul- tivariate observations, in: Proceedings of the fifth Berkeley symposium on mathematical statistics and probability, V ol. 1, Oakland, CA, USA, 1967

  38. [46]

    T. T. Cai, R. Ma, Theoretical foundations of t-sne for visualizing high-dimensional clustered data, Journal of Machine Learning Research 23 (301) (2022) 1–54

  39. [47]

    J. Lee, C. D. Campbell, S. Chen, China multi-generational panel dataset, liaoning (cmgpd-ln), 1749-1909, Data Sharing for Demographic Research (DSDR) (2010)

  40. [48]

    M. Li, S. Lu, L. Zhang, Y . Zhang, B. Zhang, A community detection method for social network based on community embedding, IEEE Trans- actions on Computational Social Systems 8 (2) (2021) 308–318

  41. [49]

    Moscatelli, J

    A. Moscatelli, J. Piquenot, M. B ´erar, P. H´eroux, S. Adam, Graph node matching for edit distance, Pattern Recognition Letters 184 (2024) 14– 20

  42. [50]

    Wills, F

    P. Wills, F. G. Meyer, Metrics for graph comparison: a practitioner’s guide, Plos one 15 (2) (2020) e0228728

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.