Pith. sign in

REVIEW 3 major objections 2 minor 29 references

When Metrics Disagree: A Meta-Analysis of Knowledge-Graph-Completion Model Benchmarking

T0 review · 3 major / 2 minor · reviewed 2026-06-27 · grok-4.3

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

Pith's one-line read Z-score aggregation resolves conflicting rank metrics by ranking DualE highest for tail prediction and FMS highest for relation prediction in KGC models.

desk verdict Z-score emerges as the balanced aggregator for conflicting KGC metrics after five tests and removals, but the abstract supplies no equations or data details to verify the Pareto claims. read the letter →

arxiv 2606.10287 v1 pith:26RZWJJK submitted 2026-06-09 cs.LG cs.CL

Haji Gul , Ajaz Ahmad Bhat This is my paper · ORCID
classification cs.LGcs.CL
keywords knowledgegraphcompletionmodelbenchmarkingmetricaggregationmulti-criteriadecisionmakingmeta-analysisZ-scoreParetooptimalityrankingconsistency
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

Standard rank metrics such as MRR and Hits@k frequently disagree on model orderings in knowledge graph completion, so the paper reframes evaluation as a multi-criteria decision problem. It tests seven aggregators on five reliability criteria—consistency, cross-dataset stability, metric independence, robustness under noise, and generalizability—each averaged over leave-one-model-out and leave-one-group-out removals. Pareto-optimal analysis identifies Z-score as the most balanced aggregator across tail and relation prediction tasks.

What carries the argument

Pareto-optimal analysis applied to seven metric aggregators evaluated on five tests (consistency, stability, independence, robustness, generalizability) using LOMO and LOGO cross-validation procedures.

What would settle it

A new KGC dataset or model set in which another aggregator produces rankings that match held-out performance better than Z-score rankings do.

Watch

Extended reading notes

Core claim

By treating KGC benchmarking as a multi-criteria decision-making task and evaluating seven aggregators through the five tests with LOMO and LOGO removals, the meta-analysis finds Z-score to be the aggregator that best balances the criteria and produces stable model rankings, with DualE leading tail prediction and FMS leading relation prediction.

Load-bearing premise

The five tests combined with LOMO and LOGO removals are sufficient and unbiased for judging which aggregator produces reliable rankings across the selected models and datasets.

Editorial extensions

If this is right

  • Model comparisons become consistent across metrics and datasets instead of depending on which metric is reported.
  • Selective reporting of favorable metrics is reduced because one aggregator is shown to dominate on the reliability tests.
  • Researchers gain evidence-based guidance for choosing an aggregator rather than defaulting to single metrics.
  • Test-sensitivity results indicate that consistency and stability tests remain stable under model removals while generalizability and independence vary most.
  • The same framework can be reused to compare future aggregators or new KGC models.

Reading between the lines

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

  • The approach could transfer to other machine-learning domains where multiple metrics produce conflicting leaderboards.
  • If Z-score rankings better predict downstream task success, they could replace ad-hoc metric selection in papers.
  • Extending the tests to include runtime or memory cost would address whether the top-ranked models remain practical.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 2 minor

Summary. The paper reframes KGC model evaluation as an MCDM problem and conducts a meta-analysis of seven aggregators evaluated on five tests (consistency, cross-dataset stability, metric independence, robustness under noise, generalizability). Tests are averaged over LOMO and LOGO removals; Pareto analysis identifies Z-score as most balanced, ranking DualE highest on tail prediction and FMS highest on relation prediction. A sensitivity analysis indicates consistency and stability are removal-invariant while generalizability and independence are more sensitive.

Significance. If the empirical results hold, the work supplies an evidence-based procedure for choosing metric aggregators in KGC benchmarking, directly addressing the documented problem of conflicting rank-based metrics (MRR, Hits@k, MR) across datasets and thereby reducing opportunities for selective reporting.

major comments (3)
  1. [§4] §4 (Tests and removals): The five tests plus LOMO/LOGO averaging are asserted to be jointly sufficient for ranking aggregators, yet no operational definitions, independence checks among the five dimensions, or power analysis for rank differences are supplied; without these the Pareto front can be sensitive to test selection.
  2. [§3.2] §3.2 (Data collection): The model and dataset collection used both to instantiate the five tests and to produce the final aggregator rankings creates circular dependence; the manuscript does not report an external hold-out collection or pre-registered protocol that would break this dependence.
  3. [§5.1] §5.1 (Pareto analysis): The claim that Z-score is Pareto-optimal rests on averaged scores whose variance across LOMO/LOGO folds is not reported; without fold-level dispersion or statistical tests it is unclear whether the reported dominance is robust or an artifact of the particular removal scheme.
minor comments (2)
  1. The abstract states conclusions but supplies neither the list of seven aggregators nor the concrete datasets; these should appear in the abstract or a prominent table.
  2. Notation for the five tests is introduced without a compact summary table that would allow readers to see at a glance which test addresses which failure mode of isolated metrics.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments, which identify important areas for clarification and strengthening. We respond point-by-point to the major comments below.

read point-by-point responses

  1. Referee: [§4] §4 (Tests and removals): The five tests plus LOMO/LOGO averaging are asserted to be jointly sufficient for ranking aggregators, yet no operational definitions, independence checks among the five dimensions, or power analysis for rank differences are supplied; without these the Pareto front can be sensitive to test selection.

    Authors: Operational definitions for each of the five tests appear in the subsections of §4 and are grounded in MCDM literature. We acknowledge that explicit pairwise independence checks among the dimensions and a formal power analysis for rank differences are not provided. In revision we will add a dedicated paragraph in §4 discussing test interdependencies, the rationale for joint sufficiency, and the practical constraints on power analysis given the model collection size. The LOMO/LOGO averaging already supplies empirical robustness across model subsets; we will emphasize this as partial mitigation of sensitivity to test selection. revision: partial

  2. Referee: [§3.2] §3.2 (Data collection): The model and dataset collection used both to instantiate the five tests and to produce the final aggregator rankings creates circular dependence; the manuscript does not report an external hold-out collection or pre-registered protocol that would break this dependence.

    Authors: The concern about circular dependence is valid. While LOMO and LOGO removals evaluate aggregator behavior on diverse subsets, they do not fully eliminate dependence on the original collection. We did not employ an external hold-out set or pre-register the protocol. In the revised manuscript we will add an explicit limitations paragraph acknowledging this issue and recommending that future work use hold-out collections or pre-registration to further validate aggregator rankings. revision: yes

  3. Referee: [§5.1] §5.1 (Pareto analysis): The claim that Z-score is Pareto-optimal rests on averaged scores whose variance across LOMO/LOGO folds is not reported; without fold-level dispersion or statistical tests it is unclear whether the reported dominance is robust or an artifact of the particular removal scheme.

    Authors: We agree that the absence of fold-level variance and statistical tests weakens the robustness claim for the Pareto front. Although the sensitivity analysis examines removal invariance for individual tests, we did not report dispersion across folds for the aggregator scores or apply statistical comparisons. In revision we will augment §5.1 with standard deviations across LOMO/LOGO folds and non-parametric tests comparing aggregator performances to substantiate that Z-score dominance is not an artifact of the removal scheme. revision: yes

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity detected

full rationale

The paper defines five explicit tests (consistency, cross-dataset stability, metric independence, robustness under noise, generalizability) and applies them via LOMO/LOGO removals to rank seven aggregators on a fixed collection of KGC models and datasets, then selects Z-score via Pareto analysis. This is an empirical comparison of aggregator behavior against independent criteria; the output ranking is not equivalent to the input definitions or data by construction, nor does any quoted step reduce via self-citation, fitted parameter, or ansatz smuggling. The derivation remains self-contained against external benchmarks of aggregator quality.

Assumptions & free parameters 0 free parameters · 1 assumptions · 0 invented entities

The central claim rests on the domain assumption that the chosen tests and removal schemes capture aggregator reliability without circular dependence on the model pool; no free parameters or invented entities are introduced.

assumptions (1)
  • domain assumption The five tests plus LOMO/LOGO removals are sufficient and unbiased measures of aggregator quality
    Invoked when the paper concludes Z-score is Pareto-optimal and removal-invariant.

how reviews work

0 comments
Cite this review

Pith. "Pith review of When Metrics Disagree: A Meta-Analysis of Knowledge-Graph-Completion Model Benchmarking." pith.science (2026). https://pith.science/paper/26RZWJJK

@misc{pith2026260610287,
  author       = {Pith},
  title        = {Pith review of: When Metrics Disagree: A Meta-Analysis of Knowledge-Graph-Completion Model Benchmarking},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/26RZWJJK}},
  note         = {Machine review of arXiv:2606.10287}
}
abstract

Evaluating Knowledge Graph Completion (KGC) models remains challenging because standard assessment relies on isolated rank-based metrics such as MRR, Hits$@$k, and Mean Rank, which often produce conflicting model orderings across datasets. A model that leads on MRR may trail on Hits@1, and strong performance on one dataset may not generalize to another. This fragmentation hinders comparison, enables selective reporting, and obscures real progress. We reframe KGC evaluation as a Multi-Criteria Decision-Making (MCDM) problem and present a meta-analysis of seven aggregators across five tests: consistency, cross-dataset stability, metric independence, robustness under noise, and generalizability. Each test is averaged over leave-one-model-out (LOMO) and leave-one-group-out (LOGO) removals so that reliability reflects aggregator behavior across diverse model subsets. Across tail $(h,r,?)$ and relation $(h,?,t)$ prediction, Pareto-optimal analysis identifies Z-score as the most balanced aggregator, which ranks DualE highest for tail prediction and FMS (Flow-Modulated Scoring) highest for relation prediction. A test-sensitivity analysis using the same removals shows that consistency and stability are largely removal-invariant, while generalizability and independence are the most sensitive. The framework resolves evaluation inconsistencies and offers evidence-based guidance for aggregator selection and model benchmarking in KGC.

Figures

Figures reproduced from arXiv: 2606.10287 by the authors.

Figure 1
Figure 1. Overview of the proposed Meta Analysis Model Benchmarking framework. (1) Knowledge graph completion (KGC) models generate link prediction results for both tail (ℎ, 𝑟, ?) and relation (ℎ, ?, 𝑡) tasks, covering diverse model families such as translational, semantic, deep learning, and graph neural network (GNN)-based approaches. (2) Multi-Criteria Decision-Making (MCDM) aggregator performance is evaluated under three … view at source ↗
Figure 2
Figure 2. (a) Consistency scores of MCDM aggregators with standard-deviation error bars (over one run with the Full-Set of 20 KGC models and 20 LOMO runs with 19 models at a time and 4 LOGO runs with one group at a time). (b) Heatmap of the average correlation between each aggregator and the base metrics across all 24 runs. Brighter cells denote stronger correlation. the three coefficients is reported as the stability error i… view at source ↗
Figure 3
Figure 3. Distribution of normalized rank variances across the five benchmark datasets [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (25 more)
Figure 5
Figure 5. Figure 5: Robustness of MCDM aggregators under noise injection, averaged over all noise levels. Error bars indicate ±1 standard deviation across noise levels. V         [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Robustness assessment with noise injection (1,000 iterations per level) and KGC method removal. Left: mean ab￾solute rank displacement (|ΔRank|, inverted so higher = more robust). Right: average correlation with the baseline ranking (higher = more robust). Aggregators …
Figure 8
Figure 8. Figure 8: Critical difference (CD) diagram showing average ranks of MCDM aggregators across five evaluation dimensions. Methods connected by horizontal lines are not significantly different (Nemenyi post-hoc test, 𝛼 = 0.05). C # #$ & $ $&  !  &  %#$ ## " ' $&   …
Figure 9
Figure 9. Figure 9: Radar plot of normalized performance scores across the five evaluation dimensions. D.1.1. Pareto Analysis To identify the most suitable MCDM aggregator for KGC evaluation, we perform a Pareto optimality analysis across all five test dimensions: consistency, stability, …
Figure 11
Figure 11. Figure 11: Relation consistency (a) Consistency test of MCDM aggregators. (b) Correlation between aggregators and base metrics. B               [PITH_FULL_IMAGE:figures/full_fig_p014_11.png]
Figure 12
Figure 12. Figure 12: Distribution of ranking variance across relation prediction datasets for different aggregators. shows the lowest consistency among the others. The reason VIKOR performs worse is the calculation method. Fig￾ure 31 (b) shows a comparison between aggregators and base met…
Figure 13
Figure 13. Figure 13: Metric independence of MCDM aggregators under leave-one-metric-family-out evaluation for relation prediction. indicating strong resistance to the removal of metrics. Over￾all, the independency test demonstrates that relation predic￾tion rankings remain highly stable u…
Figure 14
Figure 14. Figure 14: Robustness analysis under noise injection. E          [PITH_FULL_IMAGE:figures/full_fig_p016_14.png]
Figure 15
Figure 15. Figure 15: Robustness assessment with noise injection (1,000 iterations per level). Left: Mean absolute rank displacement (|ΔRank|, higher = more robust). Right: average correlation with baseline ranking (higher = more robust). In contrast, VIKOR is the most sensitive, as its co…
Figure 18
Figure 18. Figure 18: Tail Prediction KGC-method removal sensitivity across the five meta-evaluation tests. Using the Pareto-selected Z-score aggregator, [PITH_FULL_IMAGE:figures/full_fig_p017_18.png]
Figure 17
Figure 17. Figure 17: Pareto-based overall performance of MCDM aggre￾gators for relation prediction [PITH_FULL_IMAGE:figures/full_fig_p017_17.png]
Figure 19
Figure 19. Figure 19: Relation Prediction KGC-method removal sensi￾tivity across the five meta-evaluation tests. In [PITH_FULL_IMAGE:figures/full_fig_p018_19.png]
Figure 20
Figure 20. Figure 20: (a) Consistency test of MCDM aggregators. (b) Correlation between aggregators and base metrics. The consistency score for each aggregator is defined as the mean of these three correlations. To facilitate compar￾ison, we normalize the mean scores across all methods. In…
Figure 22
Figure 22. Figure 22: presents the normalized correlation with error bars representing ±1 standard deviation, sorted from highest to lowest correlation. WASPAS and VIKOR achieve the highest independency, indicating strong resistance to the removal of any single metric. The Z-score achieves…
Figure 23
Figure 23. Figure 23: Robustness analysis under noise injection. them suitable for deployment in noisy or uncertain evalu￾ation environments. For clarity, [PITH_FULL_IMAGE:figures/full_fig_p021_23.png]
Figure 24
Figure 24. Figure 24: Robustness assessment with noise injection (1,000 iterations per level). Left: Mean absolute rank displacement (|ΔRank|, higher = more robust). Right: average correlation with baseline ranking (higher = more robust). TOPSIS demonstrate superior robustness, maintaining…
Figure 28
Figure 28. Figure 28: Overall performance scores for each MCDM aggre￾gator. Z-Score achieves the highest score, indicating balanced performance across all tests. Z      [PITH_FULL_IMAGE:figures/full_fig_p023_28.png]
Figure 29
Figure 29. Figure 29: One KGC method removal overall performance. Z       !# !"        !#  #   $""          [PITH_FULL_IMAGE:figures/full_fig_p023_29.png]
Figure 27
Figure 27. Figure 27: A radar plot showing normalized performance (01 scale, higher is better) across five evaluation dimensions. The Z-score demonstrates the most balanced performance with strong scores across consistency, stability, and generalisability. by Borda, MOORA, WASPAS, and EDAS…
Figure 31
Figure 31. Figure 31: Relation consistency (a) Consistency test of MCDM aggregators. (b) Correlation between aggregators and base metrics. and Hits@3 across six benchmark datasets: FB15K, FB15K￾237, WN18, WN18RR, NELL995, and DDB14. Base results are reported in [PITH_FULL_IMAGE:figures/fu…
Figure 34
Figure 34. Figure 34: Robustness analysis under noise injection. V         [PITH_FULL_IMAGE:figures/full_fig_p025_34.png]
Figure 35
Figure 35. Figure 35: Robustness assessment with noise injection (1,000 iterations per level). Left: Mean absolute rank displacement (|ΔRank|, higher = more robust). Right: average correlation with baseline ranking (higher = more robust). procedure as tail prediction, examining whether an …
Figure 36
Figure 36. Figure 36: Generalizability analysis of MCDM aggregators under leave-one-dataset-out validation [PITH_FULL_IMAGE:figures/full_fig_p026_36.png]
Figure 38
Figure 38. Figure 38: Comparison of Z-Score and MOORA aggregators on relation prediction. Z-Score outperforms MOORA on four of five criteria, with MOORA showing competitive results only on Stability. Using the Pareto-selected Z-Score aggregator, [PITH_FULL_IMAGE:figures/full_fig_p026_38.png]
Figure 37
Figure 37. Figure 37: Pareto-based overall performance of MCDM aggre￾gators for relation prediction. overall Pareto performance, followed closely by MOORA, EDAS, and TOPSIS. This indicates that these methods main￾tain strong performance across multiple criteria rather than performing well …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 8 canonical work pages

  1. [1]

    Z. Sun, Z. H. Deng, J.-Y. Nie, J. Tang, Rotate: Knowledge graph embedding by relational rotation in complex space, in: ICLR, 2019

    2019

  2. [2]

    Y. Wang, S. Broscheit, R. Gemulla, A relational tucker decomposition for multi-relational link prediction, arXiv preprint arXiv:1902.00898 (2019)

    work page Pith review arXiv 1902

  3. [3]

    L. Yao, C. Mao, Y. Luo, Kg-bert: Bert for knowledge graph comple- tion, in: Proc. of AAAI, 2019

    2019

  4. [4]

    Y. Wei, Q. Huang, Y. Zhang, J. Kwok, Kicgpt: Large language model with knowledge in context for kgc, in: Proc. of EMNLP, 2023

    2023

  5. [5]

    H. Gul, A. G. Naim, A. A. Bhat, Muco-kgc: Multi-context-aware knowledge graph completion, in: Pacific-Asia Conference on Knowl- edge Discovery and Data Mining, Springer, 2025, pp. 3–15

    2025

  6. [6]

    H. Li, B. Yu, Y. Wei, K. Wang, R. Y. Da Xu, B. Wang, Kermit: Knowl- edge graph completion of enhanced relation modeling with inverse transformation, Knowledge-Based Systems 324 (2025) 113500

    2025

  7. [7]

    Rossi, D

    A. Rossi, D. Barbosa, D. Firmani, A. Matinata, P. Merialdo, Knowl- edge graph embedding for link prediction: A comparative analysis, TKDD (2021)

    2021

  8. [8]

    Toutanova, D

    K. Toutanova, D. Chen, Observed versus latent features for knowledge base and text inference, in: Proceedings of the 3rd workshop on continuous vector space models and their compositionality, 2015, pp. 57–66

    2015

Show all 29 references
  1. [9]

    Kadlec, O

    R. Kadlec, O. Bajgar, J. Kleindienst, Knowledge base completion: Baselines strike back, in: Proceedings of the 2nd Workshop on Representation Learning for NLP, 2017, pp. 69–74

    2017

  2. [10]

    Ruffinelli, S

    D. Ruffinelli, S. Broscheit, R. Gemulla, You can teach an old dog new tricks! on training knowledge graph embeddings (2020)

    2020

  3. [11]

    Z. Sun, Q. Zhang, W. Hu, C. Wang, M. Chen, F. Akrami, C. Li, A benchmarking study of embedding-based entity alignment for knowl- edge graphs, arXiv preprint arXiv:2003.07743 (2020)

    2003

  4. [12]

    A. Rao, N. A. Krishnan, C. R. Rivero, Using model calibration to evaluate link prediction in knowledge graphs, in: Proceedings of the ACM Web Conference 2024 (WWW ’24), Association for Computing Machinery, New York, NY, USA, 2024, pp. 2042–2051. doi:10.1145/3589334.3645506

    2024 doi

  5. [13]

    M. K. Egger, W. Ma, D. Mottin, P. Karras, I. Bordino, F. Gullo, A. Anagnostopoulos, Relik: A reliability measure for knowledge graph embeddings, in: Proceedings of the ACM Web Conference 2024 (WWW ’24), Association for Computing Machinery, New York, NY, USA, 2024. doi:10.1145/...

    2024 doi

  6. [14]

    H. Gul, A. G. Naim, A. A. Bhat, Kg-edas: A meta-metric framework for evaluating knowledge graph completion models, arXiv preprint arXiv:2508.15357 (2025). H. Gul et al.: Preprint submitted to Elsevier Page 18 of 26 When Metrics Disagree: A Meta-Analysis of KGC Model Benchmarking

    2025

  7. [15]

    J. Xu, S. Zhang, H. Xie, H. Zhang, K. Miao, Q. Fu, Knowledge graph completion based on a hierarchical graph attention network with structural information, Knowledge-Based Systems (2025) 115164

    2025

  8. [16]

    Gyani, A

    J. Gyani, A. Ahmed, M. A. Haq, Mcdm and various prioritization methods in ahp for css: A comprehensive review, IEEE Access 10 (2022) 33492–33511

    2022

  9. [17]

    E. K. Zavadskas, V . Podvezko, Integrated determination of objective criteria weights in mcdm, International Journal of Information Technology & Decision Making 15 (2016) 267–283

    2016

  10. [18]

    Z. Sun, S. Vashishth, S. Sanyal, P. Talukdar, Y. Yang, A re- evaluation of knowledge graph completion methods, in: D. Jurafsky, J. Chai, N. Schluter, J. Tetreault (Eds.), Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, Association for C...

    2020

  11. [19]

    URL: https://aclanthology.org/2020.acl-main.489/. doi: 10. 18653/v1/2020.acl-main.489

    2020

  12. [20]

    Pezeshkpour, Y

    P. Pezeshkpour, Y. Tian, S. Singh, Revisiting evaluation of knowledge base completion models, arXiv preprint arXiv:2002.00967 (2020)

    2002

  13. [21]

    Mohamed, S

    A. Mohamed, S. Parambath, Z. Kaoudi, A. Aboulnaga, Popularity agnostic evaluation of knowledge graph embeddings, in: J. Peters, D. Sontag (Eds.), Proceedings of the 36th Conference on Uncer- tainty in Artificial Intelligence (UAI), volume 124 of Proceedings of Machine Learning ...

    2020

  14. [22]

    S. Moon, Y. Ko, How sharp and bias-robust is a model? dual evalua- tion perspectives on knowledge graph completion, in: Proceedings of the Nineteenth ACM International Conference on Web Search and Data Mining, WSDM ’26, Association for Computing Machinery, New York, NY, USA, 2...

    2026 doi

  15. [23]

    Habiburahman, K

    M. Habiburahman, K. R. S. Wiharja, M. Fikriansyah, Beyond benchmarks: Assessing knowledge graph completion methods on non-benchmark employee data, in: 2024 International Conference on Data Science and Its Applications (ICODSA), volume 11, IEEE, 2024, pp. 28–33. doi: 10.1109/ic...

    2024 doi

  16. [24]

    H. Peng, L. Yao, S. Yang, Z. Zhang, J. Shi, X. Liao, H. Xiong, Causal inference-based knowledge graph completion using multi- head attention for fault diagnosis, Knowledge-Based Systems (2026) 116031

    2026

  17. [25]

    Z. Cao, C. Luo, Link prediction for knowledge graphs based on extended relational graph attention networks, Expert Syst. Appl. 259 (2025)

    2025

  18. [26]

    Z. Cao, Q. Xu, Z. Yang, X. Cao, Q. Huang, Dual quaternion knowledge graph embeddings, Proceedings of the AAAI Conference on Artificial Intelligence 35 (2021) 6894–6902

    2021

  19. [27]

    Ferrari, G

    I. Ferrari, G. Frisoni, P. Italiani, G. Moro, C. Sartori, Comprehensive analysis of knowledge graph embedding techniques benchmarked on link prediction, Electronics 11 (2022)

    2022

  20. [28]

    Rossi, D

    A. Rossi, D. Barbosa, D. Firmani, A. Matinata, P. Merialdo, Knowl- edge graph embedding for link prediction: A comparative analysis, ACM Trans. Knowl. Discov. Data 15 (2021)

    2021

  21. [29]

    Zhang, J

    Z. Zhang, J. Cai, Y. Zhang, J. Wang, Learning hierarchy-aware knowledge graph embeddings for link prediction, in: Proc. of AAAI, 2020. A. Base Results This supplementary document provides the complete baseline and ablation results for the multi-criteria decision- making (MCDM)...

    2020

Pith tools

Reviewed June 27, 2026 · model on record in the stance chip above.