Pith. sign in

REVIEW 4 major objections 3 minor 36 references

FITTER: Vocabulary-Agnostic Cross-Domain Inference on Temporal Knowledge Graphs

T0 review · 4 major / 3 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a vocabulary-agnostic structural model, FITTER, can perform temporal knowledge graph link prediction on a target graph with entirely unseen entities, relations, and timestamps, and reports consistent gains over…

desk verdict A genuinely new setting and an honest leakage appendix, but the fully-inductive claim rests on a disjoint-vocabulary protocol that the ICEWS-to-ICEWS rows likely violate. read the letter →

arxiv 2608.10668 v1 pith:LWBJDVPX submitted 2026-08-11 cs.AI

Jiaxin Pan , Mojtaba Nayyeri , Osama Mohammed , Daniel Hernandez , Rongchuan Zhang , Cheng Cheng , Steffen Staab This is my paper · ORCID
classification cs.AI
keywords temporalknowledgegraphinductivelinkpredictioncross-domaintransfervocabulary-agnosticneuralnetworksinusoidalpositionalencodingrelationinteractionembedding
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

FITTER addresses a gap in temporal knowledge graph reasoning: existing link prediction models assume the entities, relations, and timestamps seen at inference time are already known at training time, so each model is tied to one graph and one vocabulary. The paper tries to establish that a purely structure-driven model, with no dataset-specific entity, relation, or timestamp embeddings, can be trained on one temporal knowledge graph and then applied to a completely different graph without retraining. It represents time by relative snapshot order using sinusoidal positional encodings, and relations by their interaction patterns with other relations, so nothing in the representation depends on the names of entities, predicates, or dates. Across 15 cross-dataset transfer scenarios spanning six benchmarks that differ in domain, granularity, and time span, FITTER is reported to consistently outperform the inductive baselines. If this is right, a single trained model could be dropped onto a newly constructed temporal knowledge graph and rank plausible facts there, which would make cold-start and large-scale temporal reasoning substantially more practical.

What carries the argument

The machinery has three coordinated pieces. First, a vocabulary-agnostic relation interaction graph treats relations as nodes and connects them by head-to-head, head-to-tail, tail-to-head, and tail-to-tail interaction types; message passing on this graph produces query-conditioned relation representations that work for relation names never seen in training. Second, a sinusoidal temporal encoding $\mathrm{TE}(i)$ over snapshot index $i$ is time-shift invariant, meaning $\|\mathrm{TE}(\tau_2)-\mathrm{TE}(\tau_1)\|$ depends only on $\tau_2-\tau_1$, which is what lets a one-step temporal relation transfer across graphs with different timestamp vocabularies. Third, temporal-aware message passing runs on both a global entity graph and a local snapshot-window graph, fusing the two with a learned weight $\alpha$, and a multilayer perceptron scores the fused representation together with the temporal encoding. Together these components eliminate the need for any dataset-specific entity, relation, or timestamp embedding.

What would settle it

Construct two synthetic temporal knowledge graphs with identical snapshot counts, relation interactions, and recurrence statistics but different true time scales (for example, hourly versus yearly snapshots), train FITTER on one, and evaluate on the other; if transfer MRR stays flat the ordinal assumption survives, and if it drops sharply the relative-snapshot representation is not sufficient for cross-domain time.

Watch

Extended reading notes

Core claim

FITTER's central claim is that temporal link prediction can be made fully inductive by replacing absolute vocabularies with relative structure. Relation names are encoded through a relation interaction graph whose edges are structural interaction types, so a predicate is described by how it connects to other predicates rather than by a learned name embedding. Timestamps are encoded by sinusoidal positional encodings of their snapshot index, and the paper proves a time-shift invariance property: the Euclidean distance between two such encodings depends only on the difference in snapshot indices, not on the absolute values. Message passing over a global entity graph and a local windowed entity graph injects these temporal encodings into every propagation step, and the two views are fused before an MLP scores candidate quadruples. On the paper's evaluation, FITTER outperforms the compared inductive baselines on all 15 cross-domain transfer settings while recovering roughly 71-96 percent of the MRR of fully trained transductive models on in-domain benchmarks, using a model whose parameter count does not grow with the vocabulary.

Load-bearing premise

The load-bearing premise is that one snapshot of separation carries the same temporal meaning in every graph, so relative snapshot position can stand in for absolute time; if real duration or granularity affects the patterns being predicted, this encoding cannot transfer them.

Editorial extensions

If this is right

  • A model trained on one temporal knowledge graph can be applied directly to a newly built target graph with no shared entities, relations, or timestamps, enabling cold-start link prediction.
  • Training can be done on small, inexpensive graphs and transferred to large dense ones; the paper reports a model trained on the small ICEWS14 graph reaching 26.1 MRR on GDELT without retraining.
  • Because the approach is purely structural, cross-domain temporal transfer is available even when the target graph has no textual descriptions or language-model assistance.
  • The same fixed-size model handles both interpolation targets and extrapolation targets, so a single vocabulary-agnostic encoder covers settings that previously required separate transductive architectures.
  • Parameter count is independent of dataset size, so adding more temporal graphs does not require additional per-dataset embedding tables.
  • editorial extension: The ordinal snapshot encoding treats 'one snapshot apart' as the same temporal relationship whether the snapshots are 15 minutes or one year apart; a natural boundary test is to build synthetic graphs with identical snapshot counts and structure but different real durations and see whether transfer survives.
  • editorial extension: The paper's per-relation analysis suggests the temporal encoding matters most for relations whose events cluster in short windows, implying that for temporally uniform relations, structural transfer alone may be doing most of the work.
  • editorial extension: The time-shift invariance theorem implies the model can represent periodic patterns whose periods align with snapshot counts, but it cannot represent absolute elapsed time; systems that need real-duration reasoning would need an auxiliary duration channel.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 3 minor

Summary. The paper proposes FITTER, a vocabulary-agnostic structural model for temporal knowledge graph link prediction. It represents relations via ULTRA-style relation interaction graphs, encodes time via sinusoidal positional encodings over snapshot indices, and performs message passing over local and global entity graphs before scoring with an MLP. The authors claim FITTER is the first fully-inductive structural model for TKG link prediction, supporting cross-domain transfer to graphs with entirely unseen entities, relations, and timestamps. They prove a time-shift invariance property of the sinusoidal encoder and evaluate on 15 cross-dataset transfer settings across six benchmarks, reporting consistent improvements over INGRAM, ULTRA, and TRIX, plus transductive comparisons and ablations. Code, checkpoints, and datasets are released, and the appendix documents and corrects two data-leakage bugs in the GraphOracle baseline.

Significance. If the central claims are correct, FITTER would be a meaningful contribution to cold-start TKG reasoning and Semantic Web interoperability: it has a dataset-size-independent parameter count, a clean local/global temporal decomposition, and an unusually transparent appendix that identifies and fixes baseline leakage. The release of code and checkpoints is also a strength. However, the central empirical claim currently rests on three unverified pillars: strict disjointness of training and inference vocabularies, the transferability of snapshot-index time encodings across heterogeneous granularities, and the statistical stability of the reported gains. The baseline set is also limited to static-KG fully-inductive models retrained on temporal data, so the 'first fully-inductive' claim is not benchmarked against any temporal structural fully-inductive method; this is a contextual caveat rather than an error. The paper has the potential to be a solid contribution, but the reported evidence does not yet establish the headline claim.

major comments (4)
  1. [Section 6 / Table 6 / Table 1] The fully-inductive condition defined in Section 2.1, V_train ∩ V_inf = ∅, R_train ∩ R_inf = ∅, T_train ∩ T_inf = ∅, is not demonstrated for several reported settings and is likely violated. For example, the row 'Trained on ICEWS14 → ICEWS05-15' and its reverse pair both involve ICEWS subsets: ICEWS14 is the 2014 slice of ICEWS05-15, so under standard splits the entity and relation vocabularies overlap heavily and the 2014 timestamps appear in ICEWS05-15's validation and test periods. Table 6 reports only standard dataset statistics and no overlap counts or a disjoint-vocabulary preprocessing step. Unless such preprocessing was applied, these settings are at least partially transductive, and the claim of fully-inductive inference on entirely unseen entities, relations, and timestamps is not established for the settings where the paper most needs it. The authors should report exact overlap statistics for all 15 settings or rerun the experiments with explicit disjointness filters.
  2. [Section 6 / Table 1] The experimental protocol states that Q_valid on the target TKG is used to tune inference-time hyperparameters, including the local window size k and fusion weight α. This contradicts the Table 1 caption's 'without fine-tuning' and weakens the transfer claim: the model is not applied to the target graph in a fully parameter-frozen manner. More seriously, in the overlapping ICEWS settings, the target validation set contains the same entities, relations, and timestamps as the source training graph, so tuning k and α on it can leak source temporal statistics into the reported target numbers. Even for genuinely disjoint settings, choosing hyperparameters on the target validation split is a form of target adaptation and should be disclosed as such, with an analysis of how sensitive the reported conclusions are to this choice.
  3. [Section 5.1 / Appendix I, Theorem 1] The transferable time representation is based solely on the snapshot index i, not on actual timestamps. Theorem 1 proves only that the Euclidean distance between sinusoidal embeddings of two indices depends on the index difference; it says nothing about whether equal index differences in different graphs correspond to equal real time intervals. Since GDELT uses 15-minute snapshots and YAGO uses yearly snapshots, 'one snapshot apart' has entirely different temporal semantics in the two graphs. The paper calls this 'granularity-agnostic', but the ordinal encoding cannot transfer duration-dependent temporal patterns such as 'within one week' or 'after four years'. The theoretical analysis therefore does not validate cross-domain transfer; it only verifies a trigonometric identity. A controlled experiment with rescaled or relabeled timestamps would be needed to support the granularity-agnostic claim.
  4. [Section 7 / Table 1] All results are reported as single runs without error bars, confidence intervals, or significance tests. Some of the reported advantages are small relative to the likely run-to-run variance: for example, trained on ICEWS05-15 and tested on GDELT, FITTER's MRR is 18.9 versus 17.9 for both ULTRA and TRIX, and on ICEWS18 FITTER's Hits@10 is 29.7 versus 32.1 for ULTRA and TRIX, i.e., FITTER is worse. The phrase 'consistently outperforms' is not supported without multi-seed results or at least a significance analysis. The authors should report mean and standard deviation over multiple seeds, or otherwise quantify the stability of the 15-setting comparison.
minor comments (3)
  1. [Section 7.1 / Takeaway 1] The text says the time span ranges 'from just 1 month (GDELT) to 189 years (YAGO)', but Table 6 lists GDELT with Times = 366, which is consistent with one year at daily granularity, not one month; Section 6 also describes GDELT as covering one year. This internal inconsistency should be corrected.
  2. [Appendix I / Theorem 3] The proof of Theorem 3 claims that d_PE ≥ T guarantees solvability of M w = g for arbitrary g, using a Moore-Penrose right inverse. This is incorrect unless M has full row rank; d_PE ≥ T alone does not ensure that, and the Moore-Penrose inverse gives a least-squares solution, not an exact solution, when rows are dependent. The theorem needs a full-rank assumption on M or a construction of frequencies that guarantees it.
  3. [Table 2] The table reports ICL and GenTKG results as fixed regardless of training source, with only '–' for MRR; the absence of MRR and the fact that these models are evaluated only on extrapolation tasks should be stated more prominently in the main text, since it makes the comparison less direct than Table 2's layout suggests.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the cross-domain transfer claim is empirical and self-contained against external benchmarks, and the theoretical analysis is not an input to the model.

full rationale

The central claim—FITTER transfers across unseen entity, relation, and timestamp vocabularies—is an empirical result measured on six external benchmarks with held-out target test splits, and no fitted value is relabeled as a prediction. The model's components are externally sourced: relation interaction graphs from ULTRA [5], NBFNet-style message passing [36], T-ComplEx/TNT-ComplEx temporal scoring [13], and standard sinusoidal positional encodings; only the loss (Eq. 9) and the architecture glue are new, and nothing in the training objective depends on the target test facts. Theorem 1 (time-shift invariance) is a true trigonometric identity that is entailed by the definition of TE and is referenced (Appendix I) as support for transferability, but it is not used to define, train, or select the model; the load-bearing premise that ordinal snapshot position is a sufficient cross-domain time representation is assumed in Section 5.1, which is an unvalidated assumption rather than a circular step. Self-citations [26] (HGE, temporal pattern subset in Section 7.3) and [32] (TERO, candidate-generation protocol in Section 6) are evaluation details, not load-bearing derivation. Two caveats are correctness risks, not circularity: the Section 6 assertion of disjoint vocabularies ('ensuring no overlap in entities, relations, or timestamps between training and inference graphs') is unverified for the ICEWS14 to ICEWS05-15 pairs, whose standard splits plausibly share 2014 timestamps and heavily overlapping entity/relation vocabularies, and hyperparameters alpha and k tuned on the target validation split mean 'without retraining' still uses per-target validation signal; neither reduces the reported predictions to the model's inputs by construction.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central empirical claim rests mainly on the experimental protocol, not on a single derivation. The free parameters are small in number, but two of them are selected on target validation, which weakens the zero-shot framing. No new physical or symbolic entities are introduced; the temporal encoding is a feature map, not an invented entity.

free parameters (3)
  • beta, sinusoidal frequency scale = 10,000
    Chosen by hand after sensitivity analysis (Appendix G, Figure 4c). It sets how snapshot indices map to temporal frequencies and is not derived from first principles.
  • alpha, local/global fusion weight = 0.5 for ICEWS14, 0.8 for ICEWS05-15 and GDELT
    Tuned on the target validation split in fully-inductive experiments (Section 6), so cross-domain results include target-label-based selection of this scalar.
  • k, local window size = 0 for ICEWS14, 1 for ICEWS05-15 and GDELT
    Also tuned on target validation; controls the size of the local entity graph and directly affects the temporal context available at inference.
assumptions (5)
  • domain assumption Relative snapshot index is a transferable representation of time across datasets with different granularities and spans.
    Section 5.1 defines TE(i) using only the index i of the snapshot list; this conflates 15-minute, daily, and yearly intervals unless ordinal position carries the needed signal.
  • domain assumption Training and inference graphs have disjoint entities, relations, and timestamps in all 15 scenarios.
    Section 6 asserts disjointness, but the ICEWS-family datasets share vocabularies and no filtering procedure is described, so this premise may be false for several reported rows.
  • domain assumption ULTRA's relation interaction graph remains a valid vocabulary-agnostic relation encoder when applied to temporal facts and trained with the FITTER objective.
    The relational backbone is inherited without modification from ULTRA [5]; its invariance to relation names is assumed to carry over to the temporal setting.
  • ad hoc to paper The Euclidean distance property in Theorem 1 implies practical cross-domain transferability of temporal patterns.
    Theorem 1 proves a trigonometric identity about sinusoidal encodings; the leap from that identity to granularity-agnostic transfer is asserted in Section 5.1 and not proven.
  • domain assumption A scalar convex combination of local and global entity representations is sufficient to fuse short-term and long-range temporal evidence.
    Equation 7 uses a single alpha to blend the two message-passing outputs; the ablation shows sensitivity to alpha, so the sufficiency of this form is an assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FITTER: Vocabulary-Agnostic Cross-Domain Inference on Temporal Knowledge Graphs." pith.science (2026). https://pith.science/paper/LWBJDVPX

@misc{pith2026260810668,
  author       = {Pith},
  title        = {Pith review of: FITTER: Vocabulary-Agnostic Cross-Domain Inference on Temporal Knowledge Graphs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LWBJDVPX}},
  note         = {Machine review of arXiv:2608.10668}
}
read the original abstract

Temporal knowledge graphs are central to many uses of the Semantic Web, but existing completion methods assume the entities, relation names, and timestamps to be reasoned about are already known at training time, restricting each model to a single graph and vocabulary. We propose FITTER, the first fully-inductive structural model for temporal knowledge graph link prediction that supports cross-domain transfer: the inference graph may contain entirely unseen entities, relation names, and timestamps drawn from a different domain. FITTER represents each predicate by its interaction patterns with others and time through encodings of relative rather than absolute ordering; message-passing fuses local and global temporal context to produce vocabulary-agnostic embeddings. We prove the temporal encoding is time-shift invariant and evaluate FITTER on cross-domain, cross-graph transfer over six temporal knowledge graph benchmarks of diverse domains, granularities, and time spans. FITTER consistently outperforms inductive baselines without retraining, indicating that vocabulary-agnostic structural learning is a viable foundation for inference over the heterogeneous knowledge graphs of the Semantic Web.

Figures

Figures reproduced from arXiv: 2608.10668 by the authors.

Figure 1
Figure 1. Taxonomy of TKG inference settings along the vocabulary axis (trans [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Overview of FITTER. ULTRA constructs vocabulary-agnostic relation [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Relative Representation Transfer [PITH_FULL_IMAGE:figures/full_fig_p023_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: MRR and H@10 performance with different k, α and β. We report ICEWS14’s pre-training performance and keep all other hyper-parameters as the fixed best setting. H Case Study To intuitively demonstrate FITTER’s capability in generating compact and dis￾tinguishable embedd…
Figure 5
Figure 5. Figure 5: PCA visualizations of five quadruples with (China, Host a visit, ?, ?) [PITH_FULL_IMAGE:figures/full_fig_p026_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 32 canonical work pages

  1. [1]

    org/abs/2403.04782

    Cai, L., Mao, X., Zhou, Y., Long, Z., Wu, C., Lan, M.: A survey on temporal knowledge graph: Representation learning and applications (2024), https://arxiv. org/abs/2403.04782

    arXiv 2024

  2. [2]

    In: Proceedings of the 45th international ACM SIGIR conference on research and development in information retrieval

    Chen, M., Zhang, W., Zhu, Y., Zhou, H., Yuan, Z., Xu, C., Chen, H.: Meta- knowledge transfer for inductive knowledge graph embedding. In: Proceedings of the 45th international ACM SIGIR conference on research and development in information retrieval. pp. 927–937 (2022)

    work page 2022

  3. [3]

    Ding, Z., Cai, H., Wu, J., Ma, Y., Liao, R., Xiong, B., Tresp, V.: zrllm: Zero-shot relational learning on temporal knowledge graphs with large language models. In: FITTER: Vocabulary-Agnostic Cross-Domain Inference on TKGs 17 Proceedings of the 2024 Conference of the North American Chapter of the Asso- ciation for Computational Linguistics: Human Languag...

    work page 2024

  4. [4]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Du, E., Liu, S., Zhang, Y.: Graphoracle: Efficient fully-inductive knowledge graph reasoning via relation-dependency graphs. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 40, pp. 19055–19063 (2026)

    work page 2026

  5. [5]

    In: The Twelfth International Conference on Learning Representations (2024)

    Galkin, M., Yuan, X., Mostafa, H., Tang, J., Zhu, Z.: Towards foundation mod- els for knowledge graph reasoning. In: The Twelfth International Conference on Learning Representations (2024)

    2024

  6. [6]

    In: Proceedings of the 2018 conference on em- pirical methods in natural language processing

    García-Durán, A., Dumančić, S., Niepert, M.: Learning sequence encoders for tem- poral knowledge graph completion. In: Proceedings of the 2018 conference on em- pirical methods in natural language processing. pp. 4816–4821 (2018)

    work page 2018

  7. [7]

    In: IJCAI (2024)

    Gastinger, J., Meilicke, C., Errica, F., Sztyler, T., Schülke, A., Stuckenschmidt, H.: History repeats itself: A baseline for temporal knowledge graph forecasting. In: IJCAI (2024)

    work page 2024

  8. [8]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Goel, R., Kazemi, S.M., Brubaker, M., Poupart, P.: Diachronic embedding for temporal knowledge graph completion. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 34, pp. 3988–3995 (2020)

    work page 2020

Show all 36 references
  1. [9]

    In: International conference on learning representations (2020)

    Han, Z., Chen, P., Ma, Y., Tresp, V.: Explainable subgraph reasoning for fore- casting on temporal knowledge graphs. In: International conference on learning representations (2020)

    2020

  2. [10]

    Artificial intelligence194, 28–61 (2013)

    Hoffart, J., Suchanek, F.M., Berberich, K., Weikum, G.: Yago2: A spatially and temporally enhanced knowledge base from wikipedia. Artificial intelligence194, 28–61 (2013)

    2013

  3. [11]

    In: Proceedings of the 30th ACM international conference on information & knowledge management

    Jia, Z., Pramanik, S., Saha Roy, R., Weikum, G.: Complex temporal question answering on knowledge graphs. In: Proceedings of the 30th ACM international conference on information & knowledge management. pp. 792–802 (2021)

    2021

  4. [12]

    In: Proceedings of the 2020 Confer- ence on Empirical Methods in Natural Language Processing (EMNLP)

    Jin, W., Qu, M., Jin, X., Ren, X.: Recurrent event network: Autoregressive struc- ture inferenceover temporal knowledge graphs. In: Proceedings of the 2020 Confer- ence on Empirical Methods in Natural Language Processing (EMNLP). pp. 6669– 6683 (2020)

    2020

  5. [13]

    In: International Conference on Learning Representations (2020), https://openreview.net/forum?id=rke2P1BFwS

    Lacroix, T., Obozinski, G., Usunier, N.: Tensor decompositions for temporal knowl- edge base completion. In: International Conference on Learning Representations (2020), https://openreview.net/forum?id=rke2P1BFwS

    2020

  6. [14]

    Harvard Dataverse3(595), 28 (2015)

    Lautenschlager, J., Shellman, S., Ward, M.: Icews event aggregations. Harvard Dataverse3(595), 28 (2015)

    2015

  7. [15]

    In: Compan- ion proceedings of the the web conference 2018

    Leblay, J., Chekol, M.W.: Deriving validity time in knowledge graph. In: Compan- ion proceedings of the the web conference 2018. pp. 1771–1776 (2018)

    2018

  8. [16]

    In: Proceedings of the 2023 conference on empirical methods in natural language processing

    Lee, D.H., Ahrabian, K., Jin, W., Morstatter, F., Pujara, J.: Temporal knowledge graph forecasting without knowledge using in-context learning. In: Proceedings of the 2023 conference on empirical methods in natural language processing. pp. 544–557 (2023)

    2023

  9. [17]

    In: International Conference on Machine Learning

    Lee, J., Chung, C., Whang, J.J.: Ingram: Inductive knowledge graph embedding via relation graphs. In: International Conference on Machine Learning. pp. 18796– 18809. PMLR (2023)

    2023

  10. [18]

    Semantic web6(2), 167–195 (2015) 18 J

    Lehmann, J., Isele, R., Jakob, M., Jentzsch, A., Kontokostas, D., Mendes, P.N., Hellmann, S., Morsey, M., Van Kleef, P., Auer, S., et al.: Dbpedia–a large-scale, multilingual knowledge base extracted from wikipedia. Semantic web6(2), 167–195 (2015) 18 J. Pan et al

    2015

  11. [19]

    In: IJCAI

    Li, Y., Sun, S., Zhao, J.: Tirgn: Time-guided recurrent graph network with local- global historical patterns for temporal knowledge graph reasoning. In: IJCAI. pp. 2152–2158 (2022)

    2022

  12. [20]

    In: Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)

    Li,Z.,Guan,S.,Jin,X.,Peng, W.,Lyu,Y.,Zhu,Y.,Bai,L.,Li,W.,Guo,J., Cheng, X.: Complex evolutional pattern learning for temporal knowledge graph reasoning. In: Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). pp. 29...

    2022

  13. [21]

    In: Proceedings of the 46th international ACM SIGIR conference on re- search and development in information retrieval

    Liang, K., Meng, L., Liu, M., Liu, Y., Tu, W., Wang, S., Zhou, S., Liu, X.: Learn from relational correlations and periodic events for temporal knowledge graph rea- soning. In: Proceedings of the 46th international ACM SIGIR conference on re- search and development in informat...

    2023

  14. [22]

    Liao, R., Jia, X., Li, Y., Ma, Y., Tresp, V.: Gentkg: Generative forecasting on temporalknowledgegraphwithlargelanguagemodels.In:NAACL-HLT(Findings) (2024)

    2024

  15. [23]

    Advances in Neural Information Processing Systems34, 2034–2045 (2021)

    Liu, S., Grau, B., Horrocks, I., Kostylev, E.: Indigo: Gnn-based inductive knowl- edge graph completion using pair-wise encoding. Advances in Neural Information Processing Systems34, 2034–2045 (2021)

    2021

  16. [24]

    In: Proceed- ings of the AAAI conference on artificial intelligence

    Liu, Y., Ma, Y., Hildebrandt, M., Joblin, M., Tresp, V.: Tlogic: Temporal logical rules for explainable link forecasting on temporal knowledge graphs. In: Proceed- ings of the AAAI conference on artificial intelligence. vol. 36, pp. 4120–4127 (2022)

    2022

  17. [25]

    arXiv preprint arXiv:2401.06072 (2024)

    Luo, R., Gu, T., Li, H., Li, J., Lin, Z., Li, J., Yang, Y.: Chain of history: Learning and forecasting with llms for temporal knowledge graph completion. arXiv preprint arXiv:2401.06072 (2024)

    2024 arXiv

  18. [26]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Pan, J., Nayyeri, M., Li, Y., Staab, S.: Hge: embedding temporal knowledge graphs in a product space of heterogeneous geometric subspaces. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 38, pp. 8913–8920 (2024)

    2024

  19. [27]

    Neurocomputing568, 127063 (2024)

    Su, J., Ahmed, M., Lu, Y., Pan, S., Bo, W., Liu, Y.: Roformer: Enhanced trans- former with rotary position embedding. Neurocomputing568, 127063 (2024)

    2024

  20. [28]

    In: Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing

    Sun, H., Zhong, J., Ma, Y., Han, Z., He, K.: Timetraveler: Reinforcement learning for temporal knowledge graph forecasting. In: Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. pp. 8306–8319 (2021)

    2021

  21. [29]

    In: International conference on machine learning

    Teru, K., Denis, E., Hamilton, W.: Inductive relation prediction by subgraph rea- soning. In: International conference on machine learning. pp. 9448–9457. PMLR (2020)

    2020

  22. [30]

    In: international conference on machine learning

    Trivedi, R., Dai, H., Wang, Y., Song, L.: Know-evolve: Deep temporal reasoning for dynamic knowledge graphs. In: international conference on machine learning. pp. 3462–3471. PMLR (2017)

    2017

  23. [31]

    Com- munications of the ACM57(10), 78–85 (2014)

    Vrandečić, D., Krötzsch, M.: Wikidata: A free collaborative knowledgebase. Com- munications of the ACM57(10), 78–85 (2014)

    2014

  24. [32]

    In: Proceedings of the 28th International Conference on Computational Linguistics

    Xu, C., Nayyeri, M., Alkhoury, F., Yazdi, H.S., Lehmann, J.: Tero: A time-aware knowledge graph embedding via temporal rotation. In: Proceedings of the 28th International Conference on Computational Linguistics. pp. 1583–1593 (2020)

    2020

  25. [33]

    In: Proceedings of the 31st ACM International Conference on Information & Knowledge Management

    Zhang, F., Zhang, Z., Ao, X., Zhuang, F., Xu, Y., He, Q.: Along the time: Timeline- traced embedding for temporal knowledge graph completion. In: Proceedings of the 31st ACM International Conference on Information & Knowledge Management. pp. 2529–2538 (2022)

    2022

  26. [34]

    In: The Third Learning on Graphs Conference (2025) FITTER: Vocabulary-Agnostic Cross-Domain Inference on TKGs 19

    Zhang, Y., Bevilacqua, B., Galkin, M., Ribeiro, B.: Trix: A more expressive model for zero-shot domain transfer in knowledge graphs. In: The Third Learning on Graphs Conference (2025) FITTER: Vocabulary-Agnostic Cross-Domain Inference on TKGs 19

    2025

  27. [35]

    In: Pro- ceedings of the AAAI Conference on Artificial Intelligence

    Zhu, C., Chen, M., Fan, C., Cheng, G., Zhang, Y.: Learning from history: Model- ing temporal knowledge graphs with sequential copy-generation networks. In: Pro- ceedings of the AAAI Conference on Artificial Intelligence. vol. 35, pp. 4732–4740 (2021)

    2021

  28. [36]

    Advances in neural information processing systems34, 29476–29490 (2021) 20 J

    Zhu, Z., Zhang, Z., Xhonneux, L.P., Tang, J.: Neural bellman-ford networks: A general graph neural network framework for link prediction. Advances in neural information processing systems34, 29476–29490 (2021) 20 J. Pan et al. A Parameter Count and Complexity From Table 5, we ...

    2021

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.