Pith. sign in

REVIEW 4 major objections 5 minor 41 references

Evidential Spectrum-Aware Contrastive Learning for OOD Detection in Dynamic Graphs

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read EviSEC claims that out-of-distribution (OOD) detection in dynamic graphs can be recast as an uncertainty-estimation problem, with a Dirichlet posterior's uncertainty serving as the OOD score.

desk verdict Eq. 14's contrastive loss pushes negative samples toward confident wrong classes, not flat ones, contradicting the paper's central uncertainty-widening claim; otherwise a sensible EDL-for-dynamic-graphs OOD paper with extensive experiments. read the letter →

arxiv 2506.07417 v2 pith:LGI4O4TQ submitted 2025-06-09 cs.LG cs.AI

Nan Sun , Xixun Lin , Zhiheng Zhou , Yanmin Shang , Zhenlin Cheng , Yanan Cao This is my paper
classification cs.LGcs.AI
keywords dynamicgraphout-of-distributiondetectionevidentialdeeplearningspectrumcontrastiveDirichletuncertainty
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

EviSEC claims that out-of-distribution (OOD) detection in dynamic graphs can be recast as an uncertainty-estimation problem. Instead of relying on a single point prediction, the model outputs a Dirichlet posterior over class probabilities, and the uncertainty of that posterior becomes the OOD score. Because training sees no OOD data, the paper adds a spectrum-aware augmentation that deletes low-frequency Laplacian components of graph snapshots to synthesize OOD-like negatives, and a contrastive loss widens the uncertainty gap between ID snapshots and these negatives. On six real-world datasets with two OOD types, the method reports an average AUROC improvement of 24.32% over eleven baselines while keeping in-distribution F1 close to state-of-the-art performance. The reason to care is that safe deployment of dynamic graph models in security-sensitive settings hinges on flagging inputs that deviate from the training distribution, and this is a parameter-light way to learn such a flag.

What carries the argument

The load-bearing object is the evidential head: exponentiated logits minus one form a Dirichlet concentration vector $\alpha$, from which class probabilities are the mean of the posterior and uncertainty is $u = K/\alpha_{\text{sum}}$. Two evidential losses, cross-entropy on the posterior and a KL penalty toward a uniform Dirichlet, keep ID distributions sharp. The second mechanism is spectrum-aware augmentation: eigen-decomposing the normalized Laplacian, keeping all high-frequency eigenspaces, keeping only the first $r$ fraction of low-frequency eigenspaces, and reconstructing a graph $L^-_r$ that approximates an OOD sample; a contrastive loss then maximizes the separation between the uncertainty of original snapshots and augmentations. Together they convert the missing-OOD-data problem into a self-supervised widening of the uncertainty gap.

What would settle it

Run EviSEC with OOD types it never trained on, such as temporal shifts where test snapshots come from a later time period or label distribution shift without structural manipulation, and measure whether the AUROC gain over the strongest baseline persists; if the uncertainty gap learned from low-frequency deletions does not separate these real OOD samples, the core hypothesis is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that a Dirichlet-evidence output head, trained jointly with a spectrum-based contrastive loss, produces an uncertainty score that separates in-distribution from out-of-distribution dynamic graph snapshots far more sharply than post-hoc scores such as softmax confidence, energy, or plain entropy. The model preserves the argmax class probability during evidential training, so ID classification accuracy is not sacrificed; OOD detection is driven by the total Dirichlet concentration $\alpha_{\text{sum}}$, with uncertainty $u = K/\alpha_{\text{sum}}$ as the score. The spectral augmentation removes a controlled fraction of low-frequency graph Laplacian eigenspaces to manufacture negative samples, and the contrastive term pushes these negatives toward higher uncertainty. The reported result is an average AUROC gain of 24.32% over baselines and large FPR95 reductions, with near-perfect detection on two bitcoin networks.

Load-bearing premise

The method assumes that a dynamic graph made OOD-like by deleting low-frequency Laplacian components has uncertainty behavior similar to genuinely out-of-distribution dynamic graphs, so the contrastive signal transfers to unseen OOD types.

Editorial extensions

If this is right

  • If the claim holds, post-training OOD detectors on dynamic graphs can be replaced by training-time evidential heads plus spectral negatives, yielding large gains on AUROC and FPR95.
  • The Dirichlet uncertainty score is a single scalar threshold $\gamma$, so deployment can simply compare $u$ against a learned cutoff.
  • Because ID F1 is preserved, the same model serves both prediction and OOD flagging without a separate detector.
  • The spectral augmentation recipe transfers to any dynamic graph encoder that outputs node embeddings, since it operates on the Laplacian of each snapshot.

Reading between the lines

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

  • A natural test the paper leaves implicit: whether the same low-frequency deletion works when the downstream task is link prediction rather than node or edge classification, since contrastive negatives are generated per snapshot regardless of task.
  • The method suggests a general recipe: any augmentation that produces high-uncertainty synthetic points can replace OOD data, so spectral deletion is one instance of a broader uncertainty-anchored negative mining principle.
  • One could test whether using the preservation ratio $r$ as a continuous curriculum, starting with mild deletions and increasing severity, improves calibration of the uncertainty score.
  • Because the paper's OOD generation follows the static-graph framework of GNNSAFE, a direct comparison against real-world OOD examples, such as suddenly appearing fraud rings, would clarify how much of the gain is due to the augmentation matching the test generator.
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

4 major / 5 minor

Summary. The paper proposes EviSEC, an out-of-distribution (OOD) detector for dynamic graphs that combines evidential deep learning with spectrum-aware contrastive augmentation. A dynamic GCN-GRU encoder produces evidence vectors, which are converted into Dirichlet posterior concentrations; the resulting uncertainty u = K / alpha_sum is used as the OOD score. To combat score homogenization, the method removes low-frequency components of the normalized graph Laplacian to synthesize negative samples, and a contrastive loss is added to the evidential losses. Experiments on six real-world datasets with two OOD types (structure manipulation and feature interpolation) report AUROC, AUPR, and FPR95 improvements over eleven baselines, together with ablations and ID F1 checks.

Significance. The application of evidential uncertainty to OOD detection in dynamic graphs is timely, and the six-dataset evaluation with eleven baselines is substantial. The paper provides a clear problem formulation, uses standard evidential losses, and includes an ablation study that separates the contributions of the loss terms. If the central mechanism were correctly specified, the work would be a useful step for dynamic-graph OOD detection. However, as written, the contrastive objective in Sec. 4.3 does not logically enforce the claimed high-uncertainty behavior for negative samples, and the spectral augmentation pipeline is underspecified, so the reported gains are not yet attributable to the stated design.

major comments (4)
  1. [Sec. 4.3, Eq. (14)] The contrastive loss is written as L_cl = E_{p^-}[sum_i p_i log p_i^-], which is a log-likelihood (equivalently, negative cross-entropy). Since Eq. (15) is minimized, minimizing L_cl maximizes the cross-entropy between the ID class distribution p and the negative-sample distribution p^-. For an approximately one-hot p at class c, this is equivalent to minimizing E[log p_c^-], which drives p_c^- toward zero; that can be achieved by a sharp posterior concentrated on another class, which gives low uncertainty u = K / alpha_sum, not the flat, high-uncertainty posterior claimed in the same section. If the implementation instead maximizes L_cl, then p^- is pulled toward p and the uncertainty gap narrows. The manuscript must state the optimization direction and, more importantly, either change the objective to one that directly enforces high uncertainty (e.g., a KL-to-uniform or entropy penalty on p^-) or provide an analysis showing that the stationary solution of Eq. (14) is flat. As it stands, the core score-homogenization mechanism is not established.
  2. [Sec. 4.3, Eq. (13)] The augmentation step defines a matrix L_r^- as a sum of outer products of Laplacian eigenvectors, but the dynamic graph encoder in Eq. (5) consumes an adjacency matrix A_t and a feature matrix X_t. The paper does not state how L_r^- is converted into an input graph snapshot: whether the modified Laplacian is used directly in the message-passing step, or whether an adjacency matrix is reconstructed from it. This is needed both for reproducibility and for the claim that the generated matrices are 'OOD approximations' of dynamic graphs. Please specify the full augmentation pipeline, including how node-count variations and multiple temporal snapshots are handled.
  3. [Sec. 5.1 / Sec. 5.5] The paper reports that grid search determines the three key hyper-parameters (the balancing factors rho_1 and rho_2 and the preservation ratio r) but never gives the optimal values actually used for the results in Tables 2 and 3. The GitHub link also has no commit hash or configuration file. Without these values, the central quantitative claims cannot be reproduced or checked against the sensitivity analysis in Fig. 5. Please report the chosen hyper-parameters per dataset and provide a versioned code release.
  4. [Sec. 4.3 / Sec. 5.2] The spectrum-aware augmentation is justified by the heuristic that removing low-frequency Laplacian components produces OOD-like graphs, but the test OOD sets are generated by stochastic block models and feature interpolation (Sec. 5.1), not by spectral perturbation. The paper offers no analysis (spectral, embedding-level, or otherwise) showing that the augmented negatives resemble the test OOD distribution. If the augmentation heuristic is mismatched to real OOD types, the contrastive training could teach the model a signal that does not transfer. Please add a quantitative comparison of the augmented negatives and the test OOD samples, or explicitly discuss this limitation.
minor comments (5)
  1. [Abstract] The abstract uses 'EviSAC' once while the rest of the paper uses 'EviSEC'; please unify the name.
  2. [Sec. 5.2] The section header reads 'OOD Detection Perfrmance'; it should be 'Performance'.
  3. [Table 4] The ablation table is difficult to parse: the caption '✓-means the variant with random negative edge sampling augmentation' is unclear, and the row '✓ ✓ ✓ -' appears to have an extra column. Please reformat the table so each ablation variant and its configuration are unambiguous.
  4. [Sec. 5.2] The reported 'average AUROC improvement of 24.32%' is not defined; please state explicitly that the average is taken over all baselines and all datasets, and clarify whether the percentage points are absolute improvements.
  5. [Fig. 4] Figure 4 shows ID F1 scores for only four of the six datasets, while the text refers to it after a general statement about ID performance; please clarify why UCI and AS are omitted or add them.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: EviSEC's uncertainty score is learned, but no prediction reduces to a fitted input or to a self-citation by construction.

full rationale

The paper's derivation chain is self-contained and non-circular. The OOD score is the evidential uncertainty u = K / alpha_sum (Eq. 16), obtained from learned Dirichlet concentrations alpha_i = e_i + 1 (Eq. 7), where evidence e comes from the dynamic graph encoder. Training uses three losses: L_ce-edl (Eq. 10), L_kl (Eq. 11), and L_cl (Eq. 14), with L_cl acting on synthetic negatives generated by spectrum-aware augmentation (Eq. 13). The test OOD sets are generated independently by structure manipulation (stochastic block model) and feature interpolation, not by the same spectral perturbation used for training negatives, so the reported AUROC gains are not forced by construction. No equation is defined in terms of the quantity it claims to predict, and no fitted parameter is renamed as a prediction. The self-citation [31] appears only in Related Work as an example of conformal-prediction-based uncertainty methods and is not load-bearing for the proposed method. The potential sign issue in Eq. 14 is a correctness or mechanism-consistency concern, not circularity, because minimizing that loss does not make the claimed ID/OOD uncertainty gap an identity. Therefore, the paper warrants a circularity score of 0.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the standard EDL machinery, a heuristic spectrum-augmentation rule, and synthetic OOD test data. The free parameters are three hyperparameters selected by grid search. No new physical or ontological entities are introduced; the synthesized OOD approximations are data transformations rather than independent entities.

free parameters (3)
  • rho_1 = Optimal range {0.4,0.6,0.8} on AS; exact per-dataset values not reported
    Balancing factor for the KL divergence loss in Eq. 15, selected by grid search over 0.2 to 2.0. The reported OOD detection results depend on this choice.
  • rho_2 = Optimal range {0.6,0.8,1.0} on AS; exact per-dataset values not reported
    Balancing factor for the contrastive loss in Eq. 15, selected by grid search. It controls how strongly the model widens the uncertainty gap between ID and OOD approximations.
  • preservation ratio r = Grid over {0,0.2,0.4,0.6,0.8}; optimal r in [0.2,0.4]
    Controls how much low-frequency information is kept when generating negative samples in Eq. 13. The paper reports that smaller values generate more effective OOD approximations.
assumptions (4)
  • domain assumption Uniform Dirichlet prior Dirpr(p;1) with base rate beta_i = 1/K and prior weight w = K (Eq. 7)
    The method imports the standard EDL setup from [33], which assumes a uniform prior when no prior information is available. The uncertainty formula u = K / alpha_sum used as the OOD score follows from this choice.
  • ad hoc to paper Normalized Laplacian eigenvalues can be ordered 0 <= lambda_1 <= ... < 2 and split into low and high frequency at floor(N/2) (Sec. 4.3, Eqs. 12-13)
    The split point is chosen for simplicity rather than derived. The semantic claim that low frequencies encode global structure and high frequencies reflect noise is borrowed from [34] and is a heuristic assumption.
  • domain assumption Synthetic OOD test data generated by structure manipulation and feature interpolation are representative of out-of-distribution dynamic graphs (Sec. 5.1, OOD Data)
    All evaluation and the claimed generalization to OOD detection depend on this proxy. If real OOD graphs look different from SM and FI samples, the measured AUROC/FPR95 numbers would not reflect deployment behavior.
  • domain assumption The EvolveGCN-style encoder with GCN weights updated by GRUs captures the temporal dependencies of ID dynamic graph snapshots (Sec. 4.1, Eqs. 5-6)
    The method assumes this dynamic graph encoder is a sufficiently strong base model for ID representation. No theoretical guarantee or comparison across alternative encoders is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Evidential Spectrum-Aware Contrastive Learning for OOD Detection in Dynamic Graphs." pith.science (2026). https://pith.science/paper/LGI4O4TQ

@misc{pith2026250607417,
  author       = {Pith},
  title        = {Pith review of: Evidential Spectrum-Aware Contrastive Learning for OOD Detection in Dynamic Graphs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LGI4O4TQ}},
  note         = {Machine review of arXiv:2506.07417}
}
read the original abstract

Recently, Out-of-distribution (OOD) detection in dynamic graphs, which aims to identify whether incoming data deviates from the distribution of the in-distribution (ID) training set, has garnered considerable attention in security-sensitive fields. Current OOD detection paradigms primarily focus on static graphs and confront two critical challenges: i) high bias and high variance caused by single-point estimation, which makes the predictions sensitive to randomness in the data; ii) score homogenization resulting from the lack of OOD training data, where the model only learns ID-specific patterns, resulting in overall low OOD scores and a narrow score gap between ID and OOD data. To tackle these issues, we first investigate OOD detection in dynamic graphs through the lens of Evidential Deep Learning (EDL). Specifically, we propose EviSEC, an innovative and effective OOD detector via Evidential Spectrum-awarE Contrastive Learning. We design an evidential neural network to redefine the output as the posterior Dirichlet distribution, explaining the randomness of inputs through the uncertainty of distribution, which is overlooked by single-point estimation. Moreover, spectrum-aware augmentation module generates OOD approximations to identify patterns with high OOD scores, thereby widening the score gap between ID and OOD data and mitigating score homogenization. Extensive experiments on real-world datasets demonstrate that EviSAC effectively detects OOD samples in dynamic graphs.

Figures

Figures reproduced from arXiv: 2506.07417 by the authors.

Figure 1
Figure 1. Dynamic graph examples of (a) ID samples maintain consistent structures, while OOD samples exhibit deviations. (b) Examples include cases such as SM (Struc￾ture Manipulation) and FI (Feature Interpolation), highlighting structural changes and feature variations, respectively. social communities [2], and transaction records [3]. Recently, studies on dynamic graphs have gained increasing popularity, among which Dynami… view at source ↗
Figure 2
Figure 2. Heatmaps of Dirichlet distributions in 3-class classification with four groups of concentration parameters and corresponding subjective opinions. Warm (resp. cool) colors represent relatively high (resp. low) probability density values in the distribution [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. An overall illustration of the proposed method, EviSEC, which follows a two￾stage paradigm: (1) During training, original dynamic graph samples and OOD approx￾imations generated via spectrum-aware augmentation are assigned different Dirichlet distributions by the dynamic graph encoder and the EDL module—sharper for ID sam￾ples and flatter for OOD approximations. The loss function enlarges the gap in their uncertaint… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: In-Distribution performance of classification tasks with F1 scores on the y-axis. 5.5 Parameter Sensitivity(RQ4) [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: shows the hyper-parameter sensitivity of EviSEC measured by AUROC in terms of the balancing factors ρ1 and ρ2, and the preservation rate r in augmentation. Specifically, ρ1, ρ2 range from 0.2 to 2.0 with a step size of 0.2 on the AS dataset, and r ∈ {0, 0.2, 0.4, 0.6, …

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

41 extracted references · 33 canonical work pages

  1. [1]

    In: Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining

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

    2008

  2. [2]

    In: AAAI (2015), https://networkrepository.com

    Rossi, R.A., Ahmed, N.K.: The network data repository with interactive graph analytics and visualization. In: AAAI (2015), https://networkrepository.com

    work page 2015

  3. [3]

    In: Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining

    Kumar, S., Hooi, B., Makhija, D., Kumar, M., Faloutsos, C., Subrahmanian, V.: Rev2: Fraudulent user prediction in rating platforms. In: Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining. pp. 333–341 (2018)

    work page 2018

  4. [4]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Zhang, S., Chen, L., Wang, C., Li, S., Xiong, H.: Temporal graph contrastive learning for sequential recommendation. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 38, pp. 9359–9367 (2024)

    work page 2024

  5. [5]

    ACM Transactions on Intelligent Systems and Technology15(1), 1–55 (2024)

    Khoshraftar, S., An, A.: A survey on graph representation learning methods. ACM Transactions on Intelligent Systems and Technology15(1), 1–55 (2024)

    work page 2024

  6. [6]

    Artificial Intelligence in medicine23(1), 89–109 (2001)

    Kononenko, I.: Machine learning for medical diagnosis: history, state of the art and perspective. Artificial Intelligence in medicine23(1), 89–109 (2001)

    work page 2001

  7. [7]

    Yurtsever, E., Lambert, J., Carballo, A., Takeda, K.: A survey of autonomous driving:Commonpracticesandemergingtechnologies.IEEEaccess8,58443–58469 (2020)

    work page 2020

  8. [8]

    In: Proceedings of the Sixteenth ACM International Con- ference on Web Search and Data Mining

    Liu, Y., Ding, K., Liu, H., Pan, S.: Good-d: On unsupervised graph out-of- distribution detection. In: Proceedings of the Sixteenth ACM International Con- ference on Web Search and Data Mining. pp. 339–347 (2023)

    work page 2023

Show all 41 references
  1. [9]

    arXiv preprint arXiv:2302.02914 (2023)

    Wu, Q., Chen, Y., Yang, C., Yan, J.: Energy-based out-of-distribution detection for graph neural networks. arXiv preprint arXiv:2302.02914 (2023)

    2023 arXiv

  2. [10]

    In: Artificial Intelligence and Statistics

    Ranzato, M., Boureau, Y.L., Chopra, S., LeCun, Y.: A unified energy-based frame- work for unsupervised learning. In: Artificial Intelligence and Statistics. pp. 371–

  3. [11]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Song, C., Lin, X., Shen, H., Shang, Y., Cao, Y.: Uniform: Towards unified frame- work for anomaly detection on graphs. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 39, pp. 12559–12567 (2025) 16 Nan Sun et al

    2025

  4. [12]

    In: Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining

    Yu, W., Cheng, W., Aggarwal, C.C., Zhang, K., Chen, H., Wang, W.: 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 (2018)

    2018

  5. [13]

    IEEE Transactions on Knowledge and Data Engineering35(12), 12081–12094 (2021)

    Liu, Y., Pan, S., Wang, Y.G., Xiong, F., Wang, L., Chen, Q., Lee, V.C.: Anomaly detection in dynamic graphs via transformer. IEEE Transactions on Knowledge and Data Engineering35(12), 12081–12094 (2021)

    2021

  6. [14]

    In: 2021 IEEE Interna- tional Conference on Data Mining (ICDM)

    Lin, X., Cao, J., Zhang, P., Zhou, C., Li, Z., Wu, J., Wang, B.: Disentangled deep multivariate hawkes process for learning event sequences. In: 2021 IEEE Interna- tional Conference on Data Mining (ICDM). pp. 360–369. IEEE (2021)

    2021

  7. [15]

    In: Pro- ceedings of the 22nd ACM SIGKDD international conference on Knowledge dis- covery and data mining

    Grover, A., Leskovec, J.: node2vec: Scalable feature learning for networks. In: Pro- ceedings of the 22nd ACM SIGKDD international conference on Knowledge dis- covery and data mining. pp. 855–864 (2016)

    2016

  8. [16]

    In: Proceedings of the 2017 ACM on Conference on Information and Knowledge Management

    Li, J., Dani, H., Hu, X., Tang, J., Chang, Y., Liu, H.: Attributed network embed- ding for learning in a dynamic environment. In: Proceedings of the 2017 ACM on Conference on Information and Knowledge Management. pp. 387–396 (2017)

    2017

  9. [17]

    In: Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining

    Zuo, Y., Liu, G., Lin, H., Guo, J., Hu, X., Wu, J.: Embedding temporal network via neighborhood formation. In: Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining. pp. 2857–2866 (2018)

    2018

  10. [18]

    In: International Conference on Machine Learning (2024)

    Zhang, S., Zhou, C., Liu, Y., Zhang, P., Lin, X., Ma, Z.M.: Neural jump-diffusion temporal point processes. In: International Conference on Machine Learning (2024)

    2024

  11. [19]

    In: Proceedings of the AAAI conference on artificial intelli- gence

    Pareja, A., Domeniconi, G., Chen, J., Ma, T., Suzumura, T., Kanezashi, H., Kaler, T., Schardl, T., Leiserson, C.: Evolvegcn: Evolving graph convolutional networks for dynamic graphs. In: Proceedings of the AAAI conference on artificial intelli- gence. vol. 34, pp. 5363–5370 (2020)

    2020

  12. [20]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Bastos, A., Nadgeri, A., Singh, K., Suzumura, T., Singh, M.: Learnable spectral wavelets on dynamic graphs to capture global interactions. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 37, pp. 6779–6787 (2023)

    2023

  13. [21]

    13091–13092 (2022)

    Xiang, X., Huang, T., Wang, D.: Learning to evolve on dynamic graphs (student abstract).In:ProceedingsoftheAAAIConferenceonArtificialIntelligence.vol.36, pp. 13091–13092 (2022)

    2022

  14. [22]

    Advances in neural information processing systems31(2018)

    Lee, K., Lee, K., Lee, H., Shin, J.: A simple unified framework for detecting out- of-distribution samples and adversarial attacks. Advances in neural information processing systems31(2018)

    2018

  15. [23]

    arXiv preprint arXiv:1610.02136 (2016)

    Hendrycks, D., Gimpel, K.: A baseline for detecting misclassified and out-of- distribution examples in neural networks. arXiv preprint arXiv:1610.02136 (2016)

    2016 arXiv

  16. [24]

    arXiv preprint arXiv:1706.02690 (2017)

    Liang, S., Li, Y., Srikant, R.: Enhancing the reliability of out-of-distribution image detection in neural networks. arXiv preprint arXiv:1706.02690 (2017)

    2017 arXiv

  17. [25]

    Advances in neural information processing systems33, 21464–21475 (2020)

    Liu, W., Wang, X., Owens, J., Li, Y.: Energy-based out-of-distribution detection. Advances in neural information processing systems33, 21464–21475 (2020)

    2020

  18. [26]

    Information fusion76, 243–297 (2021)

    Abdar, M., Pourpanah, F., Hussain, S., Rezazadegan, D., Liu, L., Ghavamzadeh, M., Fieguth, P., Cao, X., Khosravi, A., Acharya, U.R., et al.: A review of un- certainty quantification in deep learning: Techniques, applications and challenges. Information fusion76, 243–297 (2021)

    2021

  19. [27]

    In: Forty-first International Conference on Machine Learning (2024)

    Lin, X., Zhang, W., Shi, F., Zhou, C., Zou, L., Zhao, X., Yin, D., Pan, S., Cao, Y.: Graph neural stochastic diffusion for estimating uncertainty in node classification. In: Forty-first International Conference on Machine Learning (2024)

    2024

  20. [28]

    Advances in neural information processing systems33, 12827–12836 (2020) EviSEC for OOD Detection in Dynamic Graphs 17

    Zhao, X., Chen, F., Hu, S., Cho, J.H.: Uncertainty aware semi-supervised learning on graph data. Advances in neural information processing systems33, 12827–12836 (2020) EviSEC for OOD Detection in Dynamic Graphs 17

    2020

  21. [29]

    In: International Conference on Machine Learning

    Yoon, T., Kim, H.: Uncertainty estimation by density aware evidential deep learn- ing. In: International Conference on Machine Learning. pp. 57217–57243. PMLR (2024)

    2024

  22. [30]

    In: International Conference on Machine Learning (2025)

    Zhang, S., Zhou, C., Liu, Y., Zhang, P., Lin, X., Pan, S.: Conformal anomaly detection in event sequences. In: International Conference on Machine Learning (2025)

    2025

  23. [31]

    In: Proceedings of the ACM on Web Conference 2025

    Lin, X., Cao, Y., Sun, N., Zou, L., Zhou, C., Zhang, P., Zhang, S., Zhang, G., Wu, J.: Conformal graph-level out-of-distribution detection with adaptive data augmentation. In: Proceedings of the ACM on Web Conference 2025. pp. 4755– 4765 (2025)

    2025

  24. [32]

    Jøsang, A.: Subjective logic, vol. 3. Springer (2016)

    2016

  25. [33]

    Advances in neural information processing systems31(2018)

    Sensoy, M., Kaplan, L., Kandemir, M.: Evidential deep learning to quantify classi- fication uncertainty. Advances in neural information processing systems31(2018)

    2018

  26. [34]

    Advances in Neural Information Process- ing Systems35, 2972–2983 (2022)

    Liu, N., Wang, X., Bo, D., Shi, C., Pei, J.: Revisiting graph contrastive learning from the perspective of graph spectrum. Advances in Neural Information Process- ing Systems35, 2972–2983 (2022)

    2022

  27. [35]

    In: Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining

    Leskovec, J., Kleinberg, J., Faloutsos, C.: Graphs over time: densification laws, shrinking diameters and possible explanations. In: Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining. pp. 177–187 (2005)

    2005

  28. [36]

    arXiv preprint arXiv:1908.02591 (2019)

    Weber, M., Domeniconi, G., Chen, J., Weidele, D.K.I., Bellei, C., Robinson, T., Leiserson, C.E.: Anti-money laundering in bitcoin: Experimenting with graph con- volutional networks for financial forensics. arXiv preprint arXiv:1908.02591 (2019)

    2019 arXiv

  29. [37]

    In: 2019 IEEE International Conference on Data Mining (ICDM)

    Xu, D., Cheng, W., Luo, D., Gu, Y., Liu, X., Ni, J., Zong, B., Chen, H., Zhang, X.: Adaptive neural network for node classification in dynamic networks. In: 2019 IEEE International Conference on Data Mining (ICDM). pp. 1402–1407. IEEE (2019)

    2019

  30. [38]

    The Bell system tech- nical journal28(4), 656–715 (1949)

    Shannon, C.E.: Communication theory of secrecy systems. The Bell system tech- nical journal28(4), 656–715 (1949)

    1949

  31. [39]

    In: International Conference on Ma- chine Learning

    Deng, D., Chen, G., Yu, Y., Liu, F., Heng, P.A.: Uncertainty estimation by fisher information-based evidential deep learning. In: International Conference on Ma- chine Learning. pp. 7596–7616. PMLR (2023)

    2023

  32. [40]

    In: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

    Lee, J., Kim, S., Shin, K.: Slade: Detecting dynamic anomalies in edge streams without labels via self-supervised learning. In: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. pp. 1506–1517 (2024)

    2024

  33. [41]

    In: Pro- ceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

    Guo, Y., Yang, C., Chen, Y., Liu, J., Shi, C., Du, J.: A data-centric framework to endow graph neural networks with out-of-distribution detection ability. In: Pro- ceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. pp. 638–648 (2023)

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.