Pith. sign in

REVIEW 4 major objections 5 minor 45 references

Structural-Temporal Coupling Anomaly Detection with Dynamic Graph Transformer

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read STCAD couples structural and temporal information at two levels—hand-crafted anomaly-aware features and a two-dimensional positional encoding—and reports state-of-the-art AUC and average precision on six dynamic graph benchmarks.

desk verdict The feature engineering leaks the synthetic anomaly definition into the inputs, so the reported gains are not evidence for the method; the architecture is plausible but the evaluation needs to be redone. read the letter →

arxiv 2505.08330 v1 pith:ZDQLZKRX submitted 2025-05-13 cs.LG cs.SI

Chang Zong , Yueting Zhuang , Jian Shao , Weiming Lu This is my paper · ORCID
classification cs.LGcs.SI
keywords anomalydetectiondynamicgraphsgraphtransformerstructural-temporalcouplingpositionalencodingmixedsupervision
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 claims that treating structural and temporal information as coupled, rather than independent, makes edge anomaly detection in dynamic graphs substantially more accurate. To show this, it introduces STCAD, which feeds a dynamic graph transformer two levels of features: first-level structural and temporal indicators, and second-level coupling features that measure how an edge's endpoints drift apart, change interaction degree, and change common neighbors between consecutive snapshots. The transformer is augmented with a two-dimensional positional encoding that marks each node's snapshot and its structural role relative to the candidate edge. On six benchmark datasets, with anomalies injected as edges that never appeared before, STCAD reports average gains of 3.21% in AUC and 53.67% in average precision over prior methods, with the AP gain indicating better detection of rare anomalies.

What carries the argument

The load-bearing mechanism is the two-level anomaly-aware feature encoding combined with a two-dimensional positional encoding. First-level features are standard structural and temporal indicators: PageRank, shortest-path distance, and edge lifetime. Second-level features compute, between two consecutive snapshots, the change in distance between the edge's endpoints (always 1 at the current snapshot, so a prior distance larger than 1 signals separation), the change in the sum of their degrees, and the change in the number of common neighbors. These three coupling features are hand-crafted from observed anomaly behavior. The transformer input is a sequence of these node embeddings across snapshots, and the two-dimensional positional encoding adds a temporal position (snapshot index) and a piecewise structural position (0 for the two endpoints, 1 for a common neighbor, 2 otherwise), each projected by learned weights. This lets attention weight nodes by both when and where they sit relative to the candidate edge, which is what the paper argues lets the model separate anomalies from normal edges.

What would settle it

Run STCAD on the same six datasets with a different anomaly injection—for example, edges that existed in an earlier snapshot but reappear after a long absence, or edges whose endpoints show sudden degree spikes—and check whether the reported AUC/AP advantage over TADDY and StrGNN persists. If the margin shrinks or reverses, the coupling features are tuned to the never-seen-edge definition rather than to anomalies in general.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the coupling of structural and temporal signals must happen at the input-feature level and the positional-encoding level, not only inside the embedding model. STCAD encodes each node with independent features (PageRank, shortest-path distance to the candidate edge, and edge lifetime) plus coupling features (distance change, interaction change, and common-neighbor change between consecutive snapshots), then builds an input sequence of node embeddings across snapshots. A two-dimensional positional encoding gives each node a temporal coordinate (which snapshot it appears in) and a relative structural coordinate (endpoint, common neighbor, or peripheral), letting self-attention align events across both dimensions. Finally, mixed supervision combines binary cross-entropy on the anomaly score with a self-supervised KL-divergence loss that tries to reconstruct a randomly masked node embedding from its context. The paper reports that this configuration beats previous feature-based, GNN-based, and transformer-based methods on all six datasets, and its ablations show each component—especially the coupling features and the two-dimensional encoding—contributes to the gain.

Load-bearing premise

The evaluation defines anomalies as edges that never occurred at any timestamp and samples them uniformly, so the reported gains may be specific to that synthetic anomaly type.

Editorial extensions

If this is right

  • If STCAD's results hold, dynamic edge anomaly detection improves by encoding structural-temporal coupling at both feature and positional levels, not just by stacking graph embeddings with sequence models.
  • The two-dimensional positional encoding idea—a temporal coordinate plus a relative structural coordinate—can be applied to other dynamic graph tasks, such as temporal link prediction or event forecasting.
  • The large average-precision gains suggest the method is especially useful when anomalies are rare, which is the typical setting in fraud detection and network monitoring.
  • The hand-crafted coupling features provide an interpretable signal: anomalous candidate edges are preceded by endpoints moving apart, with smaller interaction and common-neighbor changes than normal edges.
  • The method's effectiveness with only 20% of training data indicates it can be deployed in settings where labeled anomalies are scarce.

Reading between the lines

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

  • Since the anomaly definition is exactly 'an edge never seen before,' STCAD's task reduces to a form of link prediction; on real-world graphs where an anomaly can be a previously seen edge that has become suspicious, the coupling features would likely need re-calibration.
  • The three coupling features resemble classic link-prediction heuristics (shortest-path distance, degree product, common neighbors) computed as temporal deltas; one could test whether a logistic-regression model on just those features already captures much of STCAD's advantage, isolating the transformer's added value.
  • The case study's top-ranked technology combinations are presented as candidates, not validated outcomes; a prospective study that checks whether these pairs actually appear in later documents would test whether the method anticipates real emergence.
  • The 2D positional encoding's structural coordinate is discrete (0/1/2); replacing it with a continuous relative-graph-distance or a learned structural embedding could adapt the method to graphs where the peripheral/neighbor distinction is too coarse.
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

4 major / 5 minor

Summary. The paper proposes STCAD, a dynamic-graph edge anomaly detection method built around hand-crafted structural-temporal features, a transformer with two-dimensional positional encoding, and a mixed discriminative/contextual supervision objective. The authors report consistent improvements over five baselines on six benchmark datasets, with an average gain of 3.21% in AUC and 53.67% in AP, and support the result with an ablation study, a feature analysis, a parameter-sensitivity study, and a qualitative case study on emerging-technology detection.

Significance. If the reported gains are robust, the paper makes a useful contribution by showing that structural-temporal coupling features, combined with a position-aware dynamic graph transformer, can improve edge-anomaly detection on standard benchmarks. The manuscript has several strengths: an interpretable two-level feature design, a clear ablation protocol, a parameter-sensitivity analysis, a real-world case study, and a stated intention to release code and data. The central claim, however, rests on an evaluation protocol that currently has load-bearing weaknesses: model selection is performed on the test set, no variance or significance information is provided, and the anomaly-aware features were discovered from and evaluated on the same benchmark anomaly distribution. These issues substantially reduce the confidence that the reported 53.67% average AP improvement reflects a generalizable structural-temporal coupling advantage rather than an artifact of the specific synthetic anomaly protocol.

major comments (4)
  1. [Section 5.2.2] The implementation details state that the authors 'select the best result from 300 training epochs for each experiment.' If this means the epoch is chosen by looking at test-set metrics, then the numbers in Table 2 are optimistic selections over the training trajectory rather than honest estimates of generalization. A separate validation split must be used for early stopping or epoch selection, and the test-set result at the best-validation epoch should be reported. Without this, the headline comparison in Section 5.3 cannot be interpreted as a fair evaluation.
  2. [Table 2 and Section 5.3] Every reported AUC and AP value is a single number with no standard deviation, no confidence interval, and no significance test. The claim of an average improvement of 3.21% on AUC and 53.67% on AP is therefore not accompanied by any measure of variability. The authors should run each experiment over multiple random seeds and report mean plus/minus standard deviation, and for the headline claim they should include a paired statistical test across datasets or seeds.
  3. [Equation (2) and Section 5.2.2] Equation (2) defines F_dc as the distance between the central edge's endpoints in the previous snapshot minus 1, with the justification that 'the current distance is 1.' This is only true for edges that actually exist in the current snapshot. In the evaluation protocol, anomalous edges are defined as edges that never occur at any timestamp, so for those edges the current distance is not 1. As written, F_dc is not the distance-change feature it claims to be, and for the injected anomalies it may largely encode whether the candidate edge has existed before, which is exactly the synthetic anomaly definition. The equation should be corrected to use the actual current distance, and the feature should be evaluated on anomaly types that are not defined by non-existence (for example, bursty edges, rewired edges, or temporal-shift anomalies) to show that the reported gains are not an artifact of this definition.
  4. [Section 1 and Section 4.1.2] The paper states that the anomaly-aware coupling features were 'discovered with an anomaly-targeted analysis strategy with benchmark datasets' and then describes observations such as 'two nodes on an anomalous edge are usually further apart than normal ones before anomaly occurs' that were derived from the same six datasets. Evaluating features that were hand-crafted after inspecting the benchmark anomaly behavior on those same benchmarks is a form of selection bias: the comparison in Table 2 does not measure how well the features generalize to unseen anomaly patterns. The authors should evaluate on at least one held-out dynamic-graph dataset or, preferably, on a different anomaly-generation protocol, and report whether the performance advantage persists.
minor comments (5)
  1. [Equation (13)] The formula has a mismatched parenthesis: 'Sigmoid(Linear(H(eT_i,j)),' should be 'Sigmoid(Linear(H(eT_i,j))).'
  2. [Algorithm 1, line 3] The feature name F_cn is used in the algorithm pseudocode, while Equation (4) defines F_nc; please use consistent notation.
  3. [Section 5.2.2] The description of snapshot sizes is unclear: the text lists sizes for each dataset but does not explain how the snapshots are created or how many snapshots each dataset contains. This matters because the time sequence length T and the train/test split depend on the snapshot construction.
  4. [Section 6] The case study is entirely qualitative: the authors list the top five field combinations but provide no quantitative evidence (e.g., comparison with a baseline ranking, precision-at-k, or human agreement) that the identified combinations are more useful than random or frequency-based rankings.
  5. [Throughout] The symbol L is used both for the concatenation operator in Equation (6) and for the loss in Equations (14)-(16); using different symbols would avoid ambiguity.

Circularity Check

2 steps flagged · score 7.0 of 10

The key coupling feature F_dc is a direct encoding of the injected anomaly label, so the benchmark gains are substantially built into the input features.

  1. self definitional [Section 4.1.2, Eq. (2); anomaly injection in Section 5.2.2]
    "We first discover that two nodes on an anomalous edge are usually further apart than normal ones before anomaly occurs. This distance change anomaly feature can be calculated as: F_dc(v_t^k) = Dist(S_{t−∆t}, v_i^{t−∆t}, v_j^{t−∆t}) − 1, (2) ... and the current distance is 1. ... we sample the same number of edges that never occur from previous snapshots as anomalies (positive)."

    Under the paper's injection rule, anomalies are edges that never occurred at any timestamp; such an edge is necessarily absent in the previous snapshot, so Dist(S_{t−∆t}, ...) > 1 and F_dc > 0, whereas a normal edge that already existed has Dist = 1 and F_dc = 0. Thus Eq. (2) is a direct numeric encoding of the injected anomaly label, not an independent discovered pattern. Feeding this feature into every node embedding and then reporting separation on the same six datasets makes a large part of the claimed gain an artifact of the input definition.

  2. fitted input called prediction [Section 1 and Section 5.3]
    "These features are discovered with an anomaly-targeted analysis strategy with benchmark datasets. ... our proposed STCAD outperforms all baselines on six datasets with both metrics with an average increase of 3.21% on AUC and 53.67% on AP."

    The anomaly-targeted analysis was conducted on the same six benchmark datasets and the same never-occurred-edge anomaly protocol used to produce Table 2. The hand-crafted features are therefore selected to fit the test distribution, and the reported comparison measures how well the features match the benchmark's synthetic anomaly definition rather than an out-of-sample prediction. This is feature selection on the test set, so the benchmark result is partly forced by the design process.

full rationale

The paper's derivation is an empirical evaluation rather than a formal derivation, and there are no load-bearing self-citations or imported uniqueness theorems. The central circularity is definitional: the coupling feature F_dc in Eq. (2) is constructed from the previous-snapshot distance of the candidate edge minus a fixed current distance of 1. Under the paper's own anomaly injection rule, positive examples are edges that never occurred in any snapshot, so they are necessarily absent in the previous snapshot and therefore have distance > 1, while normal edges that already existed have distance exactly 1. F_dc is thus a nearly direct numeric encoding of the injected anomaly label, and because this value is attached to every node embedding in the subgraph, the transformer can separate the injected anomalies by thresholding a single input component. The separate statement that these features were discovered with an anomaly-targeted analysis strategy on benchmark datasets, followed by evaluation on the same six datasets, turns the benchmark comparison into a test of how well the handcrafted features fit the test distribution. The test-set-based model selection ('select the best result from 300 training epochs') further weakens the inference, though that is an evaluation-protocol issue rather than a definitional circularity. Overall, a substantial part of the reported 3.21% AUC and 53.67% AP improvement is built into the input features, so the central claim is partially circular, while the transformer, 2D positional encoding, and mixed-supervision components still add learnable content.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central claim rests on hand-crafted anomaly heuristics, a synthetic negative-sampling definition of anomalies, and hyperparameters selected through sensitivity analysis on the same benchmark family. No new physical or conceptual entity is introduced; the only new constructions are feature formulas and a positional encoding variant.

free parameters (6)
  • delta_t, time lag for change features = 1
    Time lag for distance, interaction, and neighbor change features; the text says 'we find that 1 is an effective option in our work' without a reported validation sweep (Equations 2 to 4).
  • context node count C = 5
    Number of context nodes in the transformer input sequence; tuned in parameter sensitivity analysis on three datasets (Section 5.6).
  • time sequence length T = 4
    Number of snapshots per sample; tuned in parameter sensitivity analysis (Section 5.6).
  • embedding dimension D = 32
    Projection and transformer hidden size; tuned in parameter sensitivity analysis (Section 5.6).
  • number of attention layers = 2
    Number of transformer layers; tuned in parameter sensitivity analysis (Section 5.6).
  • loss balancer lambda = 1.0
    Trade-off between discriminative and contextual losses; set by hand in Section 5.2.2 with no ablation reported for this value.
assumptions (6)
  • domain assumption Anomalous edges tend to be preceded by a larger shortest-path distance between their endpoints than normal edges.
    Equation (2), introduced as a 'discovered' anomaly behavior pattern on benchmark datasets and then hard-coded as a feature.
  • domain assumption Nodes of anomalous edges tend to have smaller interaction changes around them before the anomaly occurs.
    Equation (3), used as an inductive bias for the interaction-change feature.
  • domain assumption Anomalous edges tend to have fewer common-neighbor changes before the anomaly appears.
    Equation (4), used as an inductive bias for the neighbor-change feature.
  • domain assumption Edges that never occurred at any historical timestamp are a valid proxy for true anomalous edges.
    Sections 4.3.1 and 5.2.2 define all positive anomaly labels this way, for both training and testing.
  • domain assumption Dynamic graph information can be adequately summarized by PageRank, shortest-path distance, edge lifetime, and the three change features.
    Equations (1) to (4) fix the entire input representation; the model cannot see raw graph structure beyond these hand-chosen features.
  • domain assumption Masked reconstruction of the input sequence captures contextual consistency that is useful for anomaly detection.
    Equation (15) defines the contextual loss; this is a standard self-supervised assumption borrowed from other domains, not demonstrated for this task.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Structural-Temporal Coupling Anomaly Detection with Dynamic Graph Transformer." pith.science (2026). https://pith.science/paper/ZDQLZKRX

@misc{pith2026250508330,
  author       = {Pith},
  title        = {Pith review of: Structural-Temporal Coupling Anomaly Detection with Dynamic Graph Transformer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZDQLZKRX}},
  note         = {Machine review of arXiv:2505.08330}
}
read the original abstract

Detecting anomalous edges in dynamic graphs is an important task in many applications over evolving triple-based data, such as social networks, transaction management, and epidemiology. A major challenge with this task is the absence of structural-temporal coupling information, which decreases the ability of the representation to distinguish anomalies from normal instances. Existing methods focus on handling independent structural and temporal features with embedding models, which ignore the deep interaction between these two types of information. In this paper, we propose a structural-temporal coupling anomaly detection architecture with a dynamic graph transformer model. Specifically, we introduce structural and temporal features from two integration levels to provide anomaly-aware graph evolutionary patterns. Then, a dynamic graph transformer enhanced by two-dimensional positional encoding is implemented to capture both discrimination and contextual consistency signals. Extensive experiments on six datasets demonstrate that our method outperforms current state-of-the-art models. Finally, a case study illustrates the strength of our method when applied to a real-world task.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 8 canonical work pages

  1. [1]

    In: 2011 IEEE 27th international conference on data engineering, IEEE, pp 399--409, doi:10.1109/ICDE.2011.5767885

    Aggarwal CC, Zhao Y, Philip SY (2011) Outlier detection in graph streams. In: 2011 IEEE 27th international conference on data engineering, IEEE, pp 399--409, doi:10.1109/ICDE.2011.5767885

    arXiv 2011

  2. [2]

    arXiv preprint arXiv:220108236 doi:10.48550/arXiv.2201.08236

    Cai B, Xiang Y, Gao L, et al (2022) Temporal knowledge graph completion: A survey. arXiv preprint arXiv:220108236 doi:10.48550/arXiv.2201.08236

    arXiv doi:10.48550/arxiv.2201.08236 2022

  3. [3]

    In: Proceedings of the 30th ACM international conference on Information & Knowledge Management, pp 3747--3756, doi:10.1145/3459637.3481955

    Cai L, Chen Z, Luo C, et al (2021) Structural temporal graph neural networks for anomaly detection in dynamic graphs. In: Proceedings of the 30th ACM international conference on Information & Knowledge Management, pp 3747--3756, doi:10.1145/3459637.3481955

    arXiv 2021

  4. [4]

    In: European conference on computer vision, Springer, pp 213--229, doi:10.1007/978-3-030-58452-8_13

    Carion N, Massa F, Synnaeve G, et al (2020) End-to-end object detection with transformers. In: European conference on computer vision, Springer, pp 213--229, doi:10.1007/978-3-030-58452-8_13

    doi:10.1007/978-3-030-58452-8_13 2020

  5. [5]

    In: 2009 International conference on computational science and engineering, IEEE, pp 151--158, doi:10.1109/CSE.2009.439

    De Choudhury M, Sundaram H, John A, et al (2009) Social synchrony: Predicting mimicry of user actions in online social media. In: 2009 International conference on computational science and engineering, IEEE, pp 151--158, doi:10.1109/CSE.2009.439

    work page doi:10.1109/cse.2009.439 2009

  6. [6]

    arXiv preprint arXiv:201011929 doi:10.48550/arXiv.2010.11929

    Dosovitskiy A, Beyer L, Kolesnikov A, et al (2020) An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:201011929 doi:10.48550/arXiv.2010.11929

    arXiv doi:10.48550/arxiv.2010.11929 2020

  7. [7]

    arXiv preprint arXiv:211007875 doi:10.48550/arXiv.2110.07875

    Dwivedi VP, Luu AT, Laurent T, et al (2021) Graph neural networks with learnable structural and positional representations. arXiv preprint arXiv:211007875 doi:10.48550/arXiv.2110.07875

    arXiv doi:10.48550/arxiv.2110.07875 2021

  8. [9]

    In: Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies

    Gui L, Wang B, Huang Q, et al (2022) KAT : A knowledge augmented transformer for vision-and-language. In: Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Association for Computational Linguistics, pp 956--968, doi:10.18653/v1/2022.naacl-main.70

    doi:10.18653/v1/2022.naacl-main.70 2022

Show all 45 references
  1. [10]

    In: Proceedings of The Web Conference 2020, pp 2704--2710, doi:10.1145/3366423.3380027

    Hu Z, Dong Y, Wang K, et al (2020) Heterogeneous graph transformer. In: Proceedings of The Web Conference 2020, pp 2704--2710, doi:10.1145/3366423.3380027

    2020

  2. [11]

    arXiv preprint arXiv:221000881 doi:10.48550/arXiv.2210.00881

    Krenn M, Buffoni L, Coutinho B, et al (2022) Predicting the future of ai with ai: High-quality link prediction in an exponentially growing knowledge network. arXiv preprint arXiv:221000881 doi:10.48550/arXiv.2210.00881

    2022 doi arXiv

  3. [12]

    In: 2016 IEEE 16th International Conference on Data Mining (ICDM), IEEE, pp 221--230, doi:10.1109/ICDM.2016.0033

    Kumar S, Spezzano F, Subrahmanian V, et al (2016) Edge weight prediction in weighted signed networks. In: 2016 IEEE 16th International Conference on Data Mining (ICDM), IEEE, pp 221--230, doi:10.1109/ICDM.2016.0033

    2016

  4. [13]

    In: Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining, pp 333--341, doi:10.1145/3159652.3159729

    Kumar S, Hooi B, Makhija D, et al (2018) Rev2: Fraudulent user prediction in rating platforms. In: Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining, pp 333--341, doi:10.1145/3159652.3159729

    2018

  5. [14]

    arXiv preprint arXiv:190911942

    Lan Z, Chen M, Goodman S, et al (2019) Albert: A lite bert for self-supervised learning of language representations. arXiv preprint arXiv:190911942

    2019

  6. [15]

    In: European Conference on Computer Vision, Springer, pp 121--137, doi:10.1007/978-3-030-58577-8_8

    Li X, Yin X, Li C, et al (2020) Oscar: Object-semantics aligned pre-training for vision-language tasks. In: European Conference on Computer Vision, Springer, pp 121--137, doi:10.1007/978-3-030-58577-8_8

    2020 doi

  7. [16]

    arXiv preprint arXiv:190711692

    Liu Y, Ott M, Goyal N, et al (2019) Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:190711692

    2019

  8. [17]

    IEEE Transactions on Knowledge and Data Engineering doi:10.1109/TKDE.2021.3124061

    Liu Y, Pan S, Wang YG, et al (2021) Anomaly detection in dynamic graphs via transformer. IEEE Transactions on Knowledge and Data Engineering doi:10.1109/TKDE.2021.3124061

    2021

  9. [18]

    In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp 1035--1044, doi:10.1145/2939672.2939783

    Manzoor E, Milajerdi SM, Akoglu L (2016) Fast memory-efficient anomaly detection in streaming heterogeneous graphs. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp 1035--1044, doi:10.1145/2939672.2939783

    2016

  10. [19]

    arXiv preprint arXiv:210605667 doi:10.48550/arXiv.2106.05667

    Mialon G, Chen D, Selosse M, et al (2021) Graphit: Encoding graph structure in transformers. arXiv preprint arXiv:210605667 doi:10.48550/arXiv.2106.05667

    2021 doi arXiv

  11. [20]

    Social networks 31(2):155--163

    Opsahl T, Panzarasa P (2009) Clustering in weighted networks. Social networks 31(2):155--163. doi:10.1016/j.socnet.2009.02.002

    2009 doi

  12. [21]

    ACM Computing Surveys (CSUR) 54(2):1--38

    Pang G, Shen C, Cao L, et al (2021) Deep learning for anomaly detection: A review. ACM Computing Surveys (CSUR) 54(2):1--38. doi:10.1145/3439950

    2021 doi

  13. [22]

    Journal of Computational Vision and Imaging Systems 6(1):1--4

    Raisi Z, Naiel MA, Fieguth P, et al (2020) 2d positional embedding-based transformer for scene text recognition. Journal of Computational Vision and Imaging Systems 6(1):1--4. doi:10.15353/JCVIS.V6I1.3533

    2020 doi

  14. [23]

    In: Proceedings of the 2016 SIAM international conference on data mining, SIAM, pp 189--197, doi:10.1137/1.9781611974348.22

    Ranshous S, Harenberg S, Sharma K, et al (2016) A scalable approach for outlier detection in edge streams using sketch-based approximations. In: Proceedings of the 2016 SIAM international conference on data mining, SIAM, pp 189--197, doi:10.1137/1.9781611974348.22

    2016 doi

  15. [24]

    In: Twenty-ninth AAAI conference on artificial intelligence, doi:10.1609/aaai.v29i1.9277

    Rossi R, Ahmed N (2015) The network data repository with interactive graph analytics and visualization. In: Twenty-ninth AAAI conference on artificial intelligence, doi:10.1609/aaai.v29i1.9277

    2015 doi

  16. [25]

    In: Proceedings of the 13th international conference on web search and data mining, pp 519--527, doi:10.1145/3336191.3371845

    Sankar A, Wu Y, Gou L, et al (2020) Dysat: Deep neural representation learning on dynamic graphs via self-attention networks. In: Proceedings of the 13th international conference on web search and data mining, pp 519--527, doi:10.1145/3336191.3371845

    2020

  17. [26]

    IEEE Access 9:79143--79168

    Skarding J, Gabrys B, Musial K (2021) Foundations and modeling of dynamic networks using dynamic graph neural networks: A survey. IEEE Access 9:79143--79168. doi:10.1109/ACCESS.2021.3082932

    2021

  18. [27]

    In: International Conference on Machine Learning, PMLR, pp 10347--10357, doi:10.48550/arXiv.2012.12877

    Touvron H, Cord M, Douze M, et al (2021) Training data-efficient image transformers & distillation through attention. In: International Conference on Machine Learning, PMLR, pp 10347--10357, doi:10.48550/arXiv.2012.12877

    2021 doi arXiv

  19. [28]

    Advances in neural information processing systems 30

    Vaswani A, Shazeer N, Parmar N, et al (2017) Attention is all you need. Advances in neural information processing systems 30. doi:10.48550/arXiv.1706.03762

    2017 doi arXiv

  20. [29]

    Advances in neural information processing systems 32

    Yang Z, Dai Z, Yang Y, et al (2019) Xlnet: Generalized autoregressive pretraining for language understanding. Advances in neural information processing systems 32

    2019

  21. [30]

    doi:10.48550/arXiv.2106.05234

    Ying C, Cai T, Luo S, et al (2021) Do transformers really perform badly for graph representation? Advances in Neural Information Processing Systems 34:28877--28888. doi:10.48550/arXiv.2106.05234

    2021 doi arXiv

  22. [31]

    In: Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining, pp 2672--2681, doi:10.1145/3219819.3220024

    Yu W, Cheng W, Aggarwal CC, et al (2018) Netwalk: A flexible deep embedding approach for anomaly detection in dynamic networks. In: Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining, pp 2672--2681, doi:10.1145/3219819.3220024

    2018

  23. [32]

    ACM SIGCOMM Computer Communication Review 35(1):53--61

    Zhang B, Liu R, Massey D, et al (2005) Collecting the internet as-level topology. ACM SIGCOMM Computer Communication Review 35(1):53--61. doi:10.1145/1052812.1052825

    2005

  24. [33]

    In: International Joint Conference on Artificial Intelligence, doi:10.48550/arXiv.2205.04816

    Zhang J, Wang S, Chen S (2022) Reconstruction enhanced multi-view contrastive learning for anomaly detection on attributed networks. In: International Joint Conference on Artificial Intelligence, doi:10.48550/arXiv.2205.04816

    2022 doi arXiv

  25. [34]

    In: IJCAI, pp 4419--4425, doi:10.24963/ijcai.2019/614

    Zheng L, Li Z, Li J, et al (2019) Addgraph: Anomaly detection in dynamic graph using attention-based temporal gcn. In: IJCAI, pp 4419--4425, doi:10.24963/ijcai.2019/614

    2019 doi

  26. [35]

    arXiv preprint arXiv:201004159 doi:10.48550/arXiv.2010.04159

    Zhu X, Su W, Lu L, et al (2020) Deformable detr: Deformable transformers for end-to-end object detection. arXiv preprint arXiv:201004159 doi:10.48550/arXiv.2010.04159

    2020 doi arXiv

  27. [36]

    write newline

    " write newline " cite write " FUNCTION editor.postfix editor num.names #1 > "( )" "( )" if FUNCTION editor.trans.postfix editor num.names #1 > "( )" "( )" if FUNCTION trans.postfix translator num.names #1 > "( )" "( )" if FUNCTION authors.editors.reflist.apa5 'field := 'dot :...

  28. [37]

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

    ENTRY address author booktitle chapter doi edition editor eid howpublished institution journal key keywords month note number organization pages publisher school series title type url volume year eprint archive archivePrefix primaryClass adsurl adsnote version label INTEGERS o...

  29. [38]

    write newline

    " write newline "" before.all 'output.state := FUNCTION if.digit duplicate "0" = swap duplicate "1" = swap duplicate "2" = swap duplicate "3" = swap duplicate "4" = swap duplicate "5" = swap duplicate "6" = swap duplicate "7" = swap duplicate "8" = swap "9" = or or or or or or...

  30. [39]

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

    ENTRY address archive author booktitle chapter doi edition editor eid eprint howpublished institution journal key keywords month note number organization pages publisher school series title type url volume year archivePrefix primaryClass adsurl adsnote version label extra.labe...

  31. [40]

    write newline

    " write newline "" before.all 'output.state := FUNCTION add.period duplicate empty 'skip "." * add.blank if FUNCTION if.digit duplicate "0" = swap duplicate "1" = swap duplicate "2" = swap duplicate "3" = swap duplicate "4" = swap duplicate "5" = swap duplicate "6" = swap dupl...

  32. [41]

    write newline

    " write newline "" before.all 'output.state := FUNCTION output.doi doi empty skip "doi:" doi * "" * output if FUNCTION format.archive archivePrefix empty "" archivePrefix ":" * if FUNCTION format.primaryClass primaryClass empty "" " [" primaryClass * "] " * if FUNCTION format....

  33. [42]

    write newline

    " write newline "" before.all 'output.state := FUNCTION string.to.integer 't := t text.length 'k := #1 'char.num := t char.num #1 substring 's := s is.num s "." = or char.num k = not and char.num #1 + 'char.num := while char.num #1 - 'char.num := t #1 char.num substring FUNCTI...

  34. [43]

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

    ENTRY address archive author booktitle chapter edition editor eprint howpublished institution journal key keywords month note number organization pages publisher school series title type url doi volume year archivePrefix primaryClass eid adsurl adsnote version label INTEGERS o...

  35. [44]

    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 gl...

  36. [45]

    Available from:

    ENTRY address assignee author booktitle chapter cartographer day edition editor howpublished institution inventor journal key keywords month note number organization pages part publisher school series title type volume word year eprint doi url lastchecked updated archive archi...

  37. [46]

    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 gl...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.