Pith. sign in

REVIEW 4 major objections 5 minor 61 references

Towards Pattern-aware Data Augmentation for Temporal Knowledge Graph Completion

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A pattern-aware augmentation strategy lifts temporal knowledge graph completion by up to 8.7% and cuts performance variance by 22.8%.

desk verdict Booster is a plausible and genuinely new plug-in augmentation for TKG completion with a broad experimental sweep and a novel model-preference observation, but the pseudo-positive validation is recall-only, the reported recall figure is inconsistent with its own plot, and split leakage is not ruled out. read the letter →

arxiv 2501.00252 v1 pith:OCSU3PQ6 submitted 2024-12-31 cs.LG cs.DBcs.IR

Jiasheng Zhang , Deqiang Ouyang , Shuang Liang , Jie Shao This is my paper · ORCID
classification cs.LGcs.DBcs.IR
keywords temporalknowledgegraphcompletiondataaugmentationfalsenegativefilteringmodelpreferencetriadicclosurehierarchicalscoringtwo-stagetrainingembedding
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

The paper argues that poor completion performance and unstable training in temporal knowledge graph completion (TKGC) are driven by two overlooked problems: imbalanced fact distribution across entities and timestamps, and 'model preference,' where each architecture systematically favors entities with particular properties. To counter both, it proposes Booster, a plug-and-play data augmentation strategy that filters candidate false negatives by co-occurrence frequency, scores them with a hierarchical triangle-based algorithm, and fine-tunes any pre-trained TKGC model on the recovered positives, hard negatives, and model-specific hard samples. On five real-world TKGs and five backbone models, Booster reports relative MRR improvements up to 8.7%, outperforming existing graph and knowledge graph augmentation baselines by about 7.1% on average, while reducing rank variance across training runs by 22.8% on average. If these results hold, model-agnostic augmentation is a practical route to more accurate and more stable TKG completion without redesigning the underlying models.

What carries the argument

The load-bearing mechanism is a hierarchical scoring algorithm built on triadic closures in a time-irrelevant unified graph, where relation types are first anonymized to count entity triangles and then entities are anonymized to count relation triangles. Entity triangles capture how strongly three entities prefer to be mutually connected; relation triangles capture how often three relations span the same entity triangle. A candidate missing fact $(s,r,o,t)$ receives a confidence score by aggregating these global triangle scores over its recent local structure, with softmax time-aware weights that emphasize recent interactions. If the mean score over perturbed local structures exceeds the score of the observed seed fact from which the candidate was generated, the candidate is labeled a real false negative and added to training with a smooth label; otherwise it is kept as a hard negative. This scoring procedure is what lets Booster generate new samples that fit both the global semantic patterns and the local temporal trends of a TKG.

What would settle it

Take a TKG with known ground-truth missing facts (or randomly delete a held-out set of true facts), run Booster's frequency filtering and hierarchical scoring, and compute the precision of the flagged false negatives. If precision is low, the reported gains would come from denoising or regularization rather than from recovering true missing facts; a direct check is to compare Booster's recovered samples with a manually verified set of naturally absent facts.

Watch

Extended reading notes

Core claim

The central claim is that missing-but-valid facts in a temporal knowledge graph can be distinguished from true negatives by how well they complete entity and relation triangles, and that feeding these recovered facts back as training positives while flagging hard negatives improves existing TKGC models. The paper also claims to be the first to show that current TKGC methods exhibit model preferences—for example, tensor-factorization models favor frequently interacting entities while recurrent models favor recently active ones—and that such preferences cause self-training-style augmentation to reinforce a single pattern. Booster's two-stage training first pre-trains on filtered facts to avoid false negative contamination, then fine-tunes on pattern-validated false negatives, hard negatives, and low-ranked positives that expose non-preferred patterns. The reported result is consistent MRR gains across HyTE, TA, DE, TNT, and TEMP on ICEWS14, ICEWS05-15, YAGO11k, Wikidata12k, and GDELT.

Load-bearing premise

The load-bearing assumption is that a candidate fact whose perturbed triangle score beats the score of an observed seed fact is genuinely missing rather than false; the paper validates this by recall on randomly deleted facts and never measures precision on naturally missing facts.

Editorial extensions

If this is right

  • Existing TKGC models can be upgraded by wrapping them in Booster, with reported relative MRR gains of 2.5% to 8.7% across five backbones and five datasets.
  • Sparse entities and sparse timestamps gain the most, so long-tail completion performance becomes more balanced rather than uniformly shifted.
  • Because rank variance across independent training runs drops by 22.8% on average, the results of TKGC evaluations become more reproducible.
  • Static KG and temporal graph augmentation baselines, some of which can degrade performance on TKGs, should be compared against pattern-aware augmentation in future benchmarks.

Reading between the lines

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

  • Editorial inference: Booster's benefits likely depend on the true missingness rate; on graphs with few genuinely missing facts, adding pseudo-positives could degrade rather than help, so the method should be paired with a precision check.
  • Implicit extension: the same triangle-based scoring could be reused outside training augmentation, for example to rank candidate completions at inference time or to flag likely errors in TKG curation.
  • Testable extension: applying Booster to an out-of-distribution temporal split would show whether the pattern scores generalize to new relations and entity pairs, or only capture repeated historical motifs.
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 Booster, a pattern-aware data augmentation framework for temporal knowledge graph completion (TKGC). Booster first filters potential false negatives using relation-, entity-, and time-based frequency strategies (Eqs. 2-4), then scores the candidates with a hierarchical triangle-counting algorithm on a time-irrelevant unified graph (Eqs. 5-10), and finally trains a given TKGC model in two stages: pre-training on the filtered graph and fine-tuning on identified false negatives, hard negatives, and model-specific hard samples (Eqs. 11-12). Experiments on five datasets with five backbone models report MRR improvements up to 8.7%, variance reductions, ablations, and comparisons against several graph and knowledge-graph augmentation baselines.

Significance. If the claims hold, Booster would be a useful plug-and-play augmentation for TKGC, and the paper has genuine strengths: a broad experimental matrix (five backbones, five datasets), ablation and variant studies, p-values on the main comparisons, and a public code repository. The preliminary analysis of model preferences in Section 3.4 is also an interesting observation. However, the central mechanism—labeling candidate facts as positives and training on them—is validated only by recall on randomly deleted facts and by the same heuristic's own scores. No precision measure is given for the pseudo-positive set, and the possibility of split leakage is not addressed. These issues are load-bearing because they directly affect the magnitude and direction of the reported gains.

major comments (4)
  1. [Section 4.2, Figure 4] The text states that "more than 90% of removed facts can be detected," but Figure 4(a) labels 28.1% as "Not detected" for ICEWS14, implying a detection rate of about 71.9% on that dataset. The reported recall is therefore not uniformly above 90%. More importantly, recall on randomly deleted facts does not measure the precision of the final pseudo-positive set used in Eq. (12): a candidate can be detected by the filter and still be a genuine negative after scoring. Please report precision (or an estimated false-positive rate) for the pseudo-positives on naturally missing facts, using a held-out temporal split that is not used to construct the filters or triangle counts.
  2. [Section 4.3, Eq. (10)] A candidate is labeled a real false negative whenever m_f' > m_f, where both scores are produced by the same triangle-counting heuristic on the same graph. This relative threshold can admit candidates whenever the seed fact's local structure is noisy, and the perturbation/smooth-label step downweights but never removes low-confidence candidates. The paper should report how many pseudo-positives survive the threshold, what fraction are correct according to an independent signal (e.g., human annotation or facts from a future time interval not used in scoring), and how performance changes when only the highest-confidence positives are retained.
  3. [Section 4.4 and Section 5] It is never stated whether Booster's filtering and scoring operate on the training split only or on the full graph G. Since Section 5 splits each dataset into train/validation/test at 8:1:1, if validation/test facts are included in G when constructing R(r), N(e), the triangle counts, and the candidate sets, then pseudo-positives can include test answers, which would directly inflate the reported MRR gains. Please clarify this point and rerun the main tables with all augmentation computed from training facts only.
  4. [Section 4.2 and Section 5.1] The hyperparameters top-m, the sparsity threshold k, and the perturbation count k (Section 4.3) are never given values or a selection procedure; only L_r, L_e, and L_t are reported. These parameters directly control the size and composition of the pseudo-positive set and the candidate pool, so without them the experiments are not reproducible and the sensitivity of the main results to these choices is unknown.
minor comments (5)
  1. [Section 4.3, Eq. (9)] The softmax in Eq. (9) is not fully specified: it should state over which set the softmax is computed and how the time-aware weights are normalized across the relation layer.
  2. [Figure 4] The pie-chart legend is hard to parse; please give exact detection rates in the caption or a table, and correct the text to match Figure 4(a).
  3. [Table 3] The ablation study reports only Hits@1 and Hits@10, not MRR; please either include MRR or explain why it is omitted.
  4. [Section 5.1] P-values are reported but the statistical test and number of independent runs are not described; please specify both.
  5. [Throughout] There are minor spacing inconsistencies such as "TNTBooster" versus "TNT Booster" and repeated phrases like "we can see that"; a light copyedit would be helpful.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Booster's claims are empirical and benchmarked on held-out test splits; the pseudo-positive pipeline is heuristic with evaluation gaps, not a fitted prediction or self-citation chain.

full rationale

The paper does not derive a target quantity from a fitted input. Its central claims are empirical: Booster's filtering and scoring heuristics generate pseudo-positives from TKG structure, and the benefit is measured by MRR/Hits@k on held-out test splits of standard benchmarks. No equation in Section 4 is equivalent by construction to the reported improvement; Eq. 10 defines a confidence score, and Eq. 12 uses that score as a smooth label, but the final metrics are computed on test facts not used to set thresholds or hyperparameters. The 20% random-deletion experiment in Section 4.2 measures recall of the filtering strategies; it is a sanity check of coverage, not a derivation of the test-set gain. The absence of a precision measurement for naturally missing facts and the lack of an explicit statement that G is restricted to the training split are genuine evaluation risks, but they are not circularity under the hard-rule standard because no quoted reduction shows the claimed improvement is forced by the method's own definitions or by a self-citation chain. The paper contains no load-bearing self-citations and no imported uniqueness theorem, so the appropriate finding is no significant circularity.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

Booster adds no new graph-theoretic or physical entities, but it rests on five domain assumptions about what makes a missing fact true, and on six partly unreported hyperparameters. The most honest summary is that the method converts graph topological regularities into training labels without an independent precision measurement.

free parameters (6)
  • L_r = not reported (grid {1,3,5,10,20} by validation MRR)
    Relation co-occurrence time window in Eq. 2; controls which candidate false negatives are filtered.
  • L_e = not reported (grid {1,3,5,10,20} by validation MRR)
    Time window for the local entity layer in Section 4.3; controls the local structure aggregation.
  • L_t = not reported (grid {1,3,5,10,20} by validation MRR)
    Time window for time-based filtering in Eq. 4; controls how far back repeated facts are sought.
  • top-m = not reported
    Frequency cutoff for relation and entity candidate sets in Eqs. 2 and 3; never specified numerically.
  • sparsity threshold k = not reported
    Condition N(s) <= k or N(o) <= k restricts filtering to sparse local structures in Section 4.2; value not given.
  • perturbation count k = not reported
    Number of perturbed score samples in Section 4.3; also uses the symbol k, creating ambiguity with the sparsity threshold.
assumptions (5)
  • domain assumption Edges missing from a TKG but fitting frequent relation and entity co-occurrence patterns are likely true facts.
    Used to build candidate false-negative sets RN, EN, and TN in Section 4.2; this is a heuristic, not a proven property of real TKGs.
  • domain assumption Triadic closure intensity in a time-irrelevant graph is a valid proxy for the validity of a missing fact.
    Section 4.3 defines entity scores and relation scores from triangle counts and treats high scores as evidence that a candidate is a true missing fact.
  • domain assumption Removing time information (the unified graph G') preserves the semantic patterns needed for scoring.
    Section 4.3 claims G' faithfully reflects entity and relation preferences because time noise is removed; this is assumed, not demonstrated.
  • domain assumption A true fact ranked low by the pre-trained model indicates model preference rather than noise or label difficulty.
    Section 4.4 selects model-specific hard samples from low-ranked positives; if low rank is caused by noise, fine-tuning would reinforce noise.
  • domain assumption Recall on 20% randomly removed facts generalizes to naturally missing facts.
    Figure 4 evaluates detection by removing facts from G; the paper does not measure precision on naturally missing facts.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Pattern-aware Data Augmentation for Temporal Knowledge Graph Completion." pith.science (2026). https://pith.science/paper/OCSU3PQ6

@misc{pith2026250100252,
  author       = {Pith},
  title        = {Pith review of: Towards Pattern-aware Data Augmentation for Temporal Knowledge Graph Completion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OCSU3PQ6}},
  note         = {Machine review of arXiv:2501.00252}
}
read the original abstract

Predicting missing facts for temporal knowledge graphs (TKGs) is a fundamental task, called temporal knowledge graph completion (TKGC). One key challenge in this task is the imbalance in data distribution, where facts are unevenly spread across entities and timestamps. This imbalance can lead to poor completion performance or long-tail entities and timestamps, and unstable training due to the introduction of false negative samples. Unfortunately, few previous studies have investigated how to mitigate these effects. Moreover, for the first time, we found that existing methods suffer from model preferences, revealing that entities with specific properties (e.g., recently active) are favored by different models. Such preferences will lead to error accumulation and further exacerbate the effects of imbalanced data distribution, but are overlooked by previous studies. To alleviate the impacts of imbalanced data and model preferences, we introduce Booster, the first data augmentation strategy for TKGs. The unique requirements here lie in generating new samples that fit the complex semantic and temporal patterns within TKGs, and identifying hard-learning samples specific to models. Therefore, we propose a hierarchical scoring algorithm based on triadic closures within TKGs. By incorporating both global semantic patterns and local time-aware structures, the algorithm enables pattern-aware validation for new samples. Meanwhile, we propose a two-stage training approach to identify samples that deviate from the model's preferred patterns. With a well-designed frequency-based filtering strategy, this approach also helps to avoid the misleading of false negatives. Experiments justify that Booster can seamlessly adapt to existing TKGC models and achieve up to an 8.7% performance improvement.

Figures

Figures reproduced from arXiv: 2501.00252 by the authors.

Figure 1
Figure 1. An illustration of temporal knowledge graph. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. (a) Change of the 𝑟𝑎𝑛𝑘 metric during training. Each training is independently repeated four times and the color-filled part is the fluctuation range of the 𝑟𝑎𝑛𝑘 metric across four training runs. (b) The average degree of samples with different 𝑟𝑎𝑛𝑘 fluctuation ranges. (c) MRR of the TEMP model across different timestamps. The top plot displays the density function of the MRR distribution at different epochs. SD refe… view at source ↗
Figure 3
Figure 3. The conceptual illustration of the overall archi [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (8 more)
Figure 5
Figure 5. Figure 5: An example of the hierarchical scoring algorithm, [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: (a) Performance of different methods on the GDELT dataset. (b) Performance improvements of [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: (a) Performance of TNTBooster with varying hyper-parameters on the ICEWS 05-15 dataset. (b) The number of identified false negatives with different 𝐿𝑒 and 𝐿𝑟. (c) The number of identified false negatives with different 𝐿𝑒 and 𝐿𝑡 [PITH_FULL_IMAGE:figures/full_fig_p010…
Figure 8
Figure 8. Figure 8: (a) The sensitivity of model performance to the identified false negatives. (b) Performance of the TNT model with [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: (a) Throughput of the Booster framework for processing different datasets with varying 𝐿𝑒 . (b) Throughput of the Booster framework for processing different datasets with varying 𝐿𝑟.(c) Throughput of the Booster framework for processing different datasets with varying …
Figure 10
Figure 10. Figure 10: (a) MRR w.r.t. training time on the ICEWS 14 dataset. (b) Scalability of our proposed algorithm. (c) MRR variance on [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]
Figure 11
Figure 11. Figure 11: Change of 𝑟𝑎𝑛𝑘 of four test samples that are hard to be optimized by the TEMP model [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]
Figure 12
Figure 12. Figure 12: (a) Statistical study on the variance reduction of the [PITH_FULL_IMAGE:figures/full_fig_p012_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

61 extracted references · 57 canonical work pages

  1. [1]

    Antoine Bordes, Nicolas Usunier, Alberto García-Durán, Jason Weston, and Oksana Yakhnenko. 2013. Translating Embeddings for Modeling Multi-relational Data. In NIPS. 2787–2795

    work page 2013

  2. [2]

    Elizabeth Boschee, Jennifer Lautenschlager, Sean O’Brien, Steve Shellman, James Starz, and Michael Ward. 2015. ICEWS Coded Event Data . Harvard Dataverse

    work page 2015

  3. [3]

    Heng Chang, Jie Cai, and Jia Li. 2023. Knowledge Graph Completion with Counterfactual Augmentation. In The Web Conference. 2611–2620

    work page 2023

  4. [4]

    Wei Chen, Huaiyu Wan, Yuting Wu, Shuyuan Zhao, Jiayaqi Cheng, Yuxin Li, and Youfang Lin. 2024. Local-Global History-Aware Contrastive Learning for Temporal Knowledge Graph Reasoning. In ICDE. 733–746

    work page 2024

  5. [5]

    Xiangnan Chen, Wen Zhang, Zhen Yao, Mingyang Chen, and Siliang Tang

    work page

  6. [6]

    Jinhao Cui, Heyan Chai, Xu Yang, Ye Ding, Binxing Fang, and Qing Liao. 2024. SGCL: Semantic-aware Graph Contrastive Learning with Lipschitz Graph Aug- mentation. In ICDE. 3028–3041

    work page 2024

  7. [7]

    Talukdar

    Shib Sankar Dasgupta, Swayambhu Nath Ray, and Partha P. Talukdar. 2018. HyTE: Hyperplane-based Temporally aware Knowledge Graph Embedding. In EMNLP. 2001–2011

    work page 2018

  8. [8]

    Kaize Ding, Zhe Xu, Hanghang Tong, and Huan Liu. 2022. Data Augmentation for Deep Graph Learning: A Survey. SIGKDD Explor. 24, 2 (2022), 61–77

    work page 2022

Show all 61 references
  1. [9]

    Fredo Erxleben, Michael Günther, Markus Krötzsch, Julian Mendez, and Denny Vrandecic. 2014. Introducing Wikidata to the Linked Data Web. In ISWC. 50–65

    2014

  2. [10]

    Alberto García-Durán, Sebastijan Dumancic, and Mathias Niepert. 2018. Learning Sequence Encoders for Temporal Knowledge Graph Completion. In EMNLP. 4816–4821

    2018

  3. [11]

    Schoenholz, Patrick F

    Justin Gilmer, Samuel S. Schoenholz, Patrick F. Riley, Oriol Vinyals, and George E. Dahl. 2017. Neural Message Passing for Quantum Chemistry. In ICML. 1263– 1272

    2017

  4. [12]

    Brubaker, and Pascal Poupart

    Rishab Goel, Seyed Mehran Kazemi, Marcus A. Brubaker, and Pascal Poupart

  5. [13]

    Pankaj Gupta, Venu Satuluri, Ajeet Grewal, Siva Gurumurthy, Volodymyr Zhabiuk, Quannan Li, and Jimmy Lin. 2014. Real-Time Twitter Recommen- dation: Online Motif Detection in Large Dynamic Graphs. Proc. VLDB Endow. 7, 13 (2014), 1379–1380

    2014

  6. [14]

    Xiaotian Han, Zhimeng Jiang, Ninghao Liu, and Xia Hu. 2022. G-Mixup: Graph Data Augmentation for Graph Classification. In ICML, Vol. 162. 8230–8248

    2022

  7. [15]

    Zhen Han, Peng Chen, Yunpu Ma, and Volker Tresp. 2021. Explainable Subgraph Reasoning for Forecasting on Temporal Knowledge Graphs. In ICLR

    2021

  8. [16]

    Prachi Jain, Sushant Rathi, Mausam, and Soumen Chakrabarti. 2020. Temporal Knowledge Base Completion: New Algorithms and Evaluation Protocols. In EMNLP. 3733–3747

    2020

  9. [17]

    Mingxuan Ju, Tong Zhao, Wenhao Yu, Neil Shah, and Yanfang Ye. 2023. Graph- Patcher: Mitigating Degree Bias for Graph Neural Networks via Test-time Aug- mentation. In NeurIPS

    2023

  10. [18]

    Hidetaka Kamigaito and Katsuhiko Hayashi. 2022. Comprehensive Analysis of Negative Sampling in Knowledge Graph Representation Learning. In ICML, Vol. 162. 10661–10675

    2022

  11. [19]

    Timothée Lacroix, Guillaume Obozinski, and Nicolas Usunier. 2020. Tensor Decompositions for Temporal Knowledge Base Completion. In ICLR

    2020

  12. [20]

    Kalev Leetaru and Philip A. Schrodt. 2013. Global Database of Events, Language and Tone. In ISA

    2013

  13. [21]

    Zixuan Li, Xiaolong Jin, Wei Li, Saiping Guan, Jiafeng Guo, Huawei Shen, Yuanzhuo Wang, and Xueqi Cheng. 2021. Temporal Knowledge Graph Rea- soning Based on Evolutional Representation Learning. In SIGIR. 408–417

    2021

  14. [22]

    Hongrui Liu, Binbin Hu, Xiao Wang, Chuan Shi, Zhiqiang Zhang, and Jun Zhou

  15. [23]

    Bin Lu, Ze Zhao, Xiaoying Gan, Shiyu Liang, Luoyi Fu, Xinbing Wang, and Chenghu Zhou. 2024. Graph Out-of-Distribution Generalization With Control- lable Data Augmentation. IEEE Trans. Knowl. Data Eng. 36, 11 (2024), 6317–6329

    2024

  16. [24]

    Agnes Lydia and F

    A. Agnes Lydia and F. Sagayaraj Francis. 2019. Adagrad - An optimizer for stochastic gradient descent. Int. J. Inf. Comput. Sci. 6, 5 (2019), 566–568

    2019

  17. [25]

    Tiroshan Madushanka and Ryutaro Ichise. 2024. Negative Sampling in Knowl- edge Graph Representation Learning: A Review. CoRR abs/2402.19195 (2024)

    2024 arXiv

  18. [26]

    Saurav Manchanda. 2023. Metapath-Guided Data-Augmentation For Knowledge Graphs. In CIKM. 4175–4179

    2023

  19. [27]

    Zara Nasar, Syed Waqar Jaffry, and Muhammad Kamran Malik. 2022. Named Entity Recognition and Relation Extraction: State-of-the-Art. ACM Comput. Surv. 54, 1 (2022), 20:1–20:39

    2022

  20. [28]

    Harry Shomer, Wei Jin, Wentao Wang, and Jiliang Tang. 2023. Toward Degree Bias in Embedding-Based Knowledge Graph Completion. In The Web Conference. 705–715

    2023

  21. [29]

    Suchanek, Gjergji Kasneci, and Gerhard Weikum

    Fabian M. Suchanek, Gjergji Kasneci, and Gerhard Weikum. 2007. Yago: a core of semantic knowledge. In WWW. 697–706

    2007

  22. [30]

    Yongduo Sui, Qitian Wu, Jiancan Wu, Qing Cui, Longfei Li, Jun Zhou, Xiang Wang, and Xiangnan He. 2023. Unleashing the Power of Graph Data Augmenta- tion on Covariate Distribution Shift. In NeurIPS

    2023

  23. [31]

    Jizhi Tang, Yansong Feng, and Dongyan Zhao. 2019. Learning to Update Knowl- edge Graphs by Reading News. In EMNLP-IJCNLP. 2632–2641

    2019

  24. [32]

    Xing Tang, Ling Chen, Hongyu Shi, and Dandan Lyu. 2024. DHyper: A Recurrent Dual Hypergraph Neural Network for Event Prediction in Temporal Knowledge Graphs. ACM Trans. Inf. Syst. 42, 5 (2024), 129:1–129:23

    2024

  25. [33]

    Ag- garwal, Prasenjit Mitra, and Suhang Wang

    Xianfeng Tang, Huaxiu Yao, Yiwei Sun, Yiqi Wang, Jiliang Tang, Charu C. Ag- garwal, Prasenjit Mitra, and Suhang Wang. 2020. Investigating and Mitigating Degree-Related Biases in Graph Convoltuional Networks. In CIKM. 1435–1444

    2020

  26. [34]

    Zhenwei Tang, Shichao Pei, Zhao Zhang, Yongchun Zhu, Fuzhen Zhuang, Robert Hoehndorf, and Xiangliang Zhang. 2022. Positive-Unlabeled Learning with Adversarial Data Augmentation for Knowledge Graph Completion. In IJCAI. 2248–2254

    2022

  27. [35]

    Yuxing Tian, Aiwen Jiang, Qi Huang, Jian Guo, and Yiyan Qi. 2024. Latent Diffusion-based Data Augmentation for Continuous-Time Dynamic Graph Model. In SIGKDD. 2900–2911

    2024

  28. [36]

    Jiapu Wang, Boyue Wang, Junbin Gao, Xiaoyan Li, Yongli Hu, and Baocai Yin

  29. [37]

    Jiapu Wang, Boyue Wang, Junbin Gao, Shirui Pan, Tengfei Liu, Baocai Yin, and Wen Gao. 2024. MADE: Multicurvature Adaptive Embedding for Temporal Knowledge Graph Completion. IEEE Trans. Cybern. 54, 10 (2024), 5818–5831

    2024

  30. [38]

    Jiapu Wang, Boyue Wang, Meikang Qiu, Shirui Pan, Bo Xiong, Heng Liu, Linhao Luo, Tengfei Liu, Yongli Hu, Baocai Yin, and Wen Gao. 2023. A Survey on Temporal Knowledge Graph Completion: Taxonomy, Progress, and Prospects. CoRR abs/2308.02457 (2023)

    2023 arXiv

  31. [39]

    Yiwei Wang, Yujun Cai, Yuxuan Liang, Henghui Ding, Changhu Wang, Siddharth Bhatia, and Bryan Hooi. 2021. Adaptive Data Augmentation on Temporal Graphs. In NeurIPS. 1440–1452

    2021

  32. [40]

    Hamilton

    Jiapeng Wu, Meng Cao, Jackie Chi Kit Cheung, and William L. Hamilton. 2020. TeMP: Temporal Message Passing for Temporal Knowledge Graph Completion. In EMNLP. 5730–5746

    2020

  33. [41]

    Yaochen Xie, Zhao Xu, Jingtun Zhang, Zhengyang Wang, and Shuiwang Ji. 2023. Self-Supervised Learning of Graph Neural Networks: A Unified Review. IEEE Trans. Pattern Anal. Mach. Intell. 45, 2 (2023), 2412–2429

    2023

  34. [42]

    Hao Xin and Lei Chen. 2024. KartGPS: Knowledge Base Update with Temporal Graph Pattern-based Semantic Rules. In ICDE. 5075–5087

    2024

  35. [43]

    Siheng Xiong, Yuan Yang, Faramarz Fekri, and James Clayton Kerce. 2023. TILP: Differentiable Learning of Temporal Logical Rules on Knowledge Graphs. In ICLR

    2023

  36. [44]

    Chengjin Xu, Yung-Yu Chen, Mojtaba Nayyeri, and Jens Lehmann. 2021. Tem- poral Knowledge Graph Completion using a Linear Temporal Regularizer and Multivector Embeddings. In NAACL-HLT. 2569–2578

    2021

  37. [45]

    Chenjin Xu, Mojtaba Nayyeri, Fouad Alkhoury, Hamed Shariat Yazdi, and Jens Lehmann. 2020. Temporal Knowledge Graph Completion Based on Time Series Gaussian Embedding. In ISWC. 654–671

    2020

  38. [46]

    Chengjin Xu, Mojtaba Nayyeri, Fouad Alkhoury, Hamed Shariat Yazdi, and Jens Lehmann. 2020. TeRo: A Time-aware Knowledge Graph Embedding via Temporal Rotation. In COLING. 1583–1593

    2020

  39. [47]

    Yi Xu, Junjie Ou, Hui Xu, and Luoyi Fu. 2023. Temporal Knowledge Graph Reasoning with Historical Contrastive Learning. In AAAI. 4765–4773

    2023

  40. [48]

    Jinfa Yang, Xianghua Ying, Yongjie Shi, and Bowei Xing. 2024. Tensor decompo- sitions for temporal knowledge graph completion with time perspective. Expert Syst. Appl. 237, Part A (2024), 121267

    2024

  41. [49]

    Naimeng Yao, Qing Liu, Yi Yang, Weihua Li, and Quan Bai. 2023. Entity-Relation Distribution-Aware Negative Sampling for Knowledge Graph Embedding. In ISWC. 234–252

    2023

  42. [50]

    Yuanzhou Yao, Zhao Zhang, Yongjun Xu, and Chao Li. 2022. Data Augmentation for Few-Shot Knowledge Graph Completion from Hierarchical Perspective. In COLING. 2494–2503

    2022

  43. [51]

    Fu Zhang, Hongzhi Chen, Yuzhe Shi, Jingwei Cheng, and Jinghao Lin. 2024. Joint framework for tensor decomposition-based temporal knowledge graph completion. Inf. Sci. 654 (2024), 119853

    2024

  44. [52]

    Mengqi Zhang, Yuwei Xia, Qiang Liu, Shu Wu, and Liang Wang. 2023. Learning Long- and Short-term Representations for Temporal Knowledge Graph Reason- ing. In The Web Conference. 2412–2422

    2023

  45. [53]

    Qianru Zhang, Lianghao Xia, Xuheng Cai, Siu-Ming Yiu, Chao Huang, and Christian S. Jensen. 2024. Graph Augmentation for Recommendation. In ICDE. 13 557–569

    2024

  46. [54]

    Shuaicheng Zhang, Yada Zhu, and Dawei Zhou. 2023. TGEditor: Task-Guided Graph Editing for Augmenting Temporal Financial Transaction Networks. In ICAIF. 219–226

    2023

  47. [55]

    Yongqi Zhang, Quanming Yao, and Lei Chen. 2021. Simple and automated negative sampling for knowledge graph embedding. VLDB J. 30, 2 (2021), 259– 285

    2021

  48. [56]

    Yuyue Zhao, Xiang Wang, Jiawei Chen, Yashen Wang, Wei Tang, Xiangnan He, and Haiyong Xie. 2023. Time-aware Path Reasoning on Knowledge Graph for Recommendation. ACM Trans. Inf. Syst. 41, 2 (2023), 26:1–26:26

    2023

  49. [57]

    Xinyi Zhu, Liping Wang, Hao Xin, Xiaohan Wang, Zhifeng Jia, Jiyao Wang, Chunming Ma, and Yuxiang Zengt. 2023. T-FinKB: A Platform of Temporal Financial Knowledge Base Construction. In ICDE. 3671–3674. 14

    2023

  50. [2020]

    Diachronic Embedding for Temporal Knowledge Graph Completion. In AAAI. 3988–3995

  51. [2022]

    In The Web Conference

    Confidence May Cheat: Self-Training on Graph Neural Networks under Distribution Shift. In The Web Conference. 1248–1258

  52. [2023]

    Negative Sampling with Adaptive Denoising Mixup for Knowledge Graph Embedding. In ISWC. 253–270

  53. [2024]

    IEEE Trans

    QDN: A Quadruplet Distributor Network for Temporal Knowledge Graph Completion. IEEE Trans. Neural Networks Learn. Syst. 35, 10 (2024), 14018–14030

    2024

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.