The DOI matches its Crossref record; 23 of the 23 references it cites were found in a registry or an index; 5 of the 14 sources its claims cite were checked for support.
TuckER: Tensor Factorization for Knowledge Graph Completion
Open accessAbstract
Ivana Balazevic, Carl Allen, Timothy Hospedales. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
TuckER is a linear link-prediction model based on Tucker decomposition of the binary knowledge-graph tensor, using a shared core tensor and entity/relation embeddings. It achieves state-of-the-art results on standard datasets, is fully expressive, and subsumes several prior linear models as special cases.
Key findings
- TuckER reaches state-of-the-art link prediction results on WN18RR (MRR .470, Hits@10 .526, Hits@3 .482, Hits@1 .443) and FB15k-237 (MRR .358, Hits@10 .544, Hits@3 .394, Hits@1 .266), and on WN18 (MRR .953, Hits@10 .958, Hits@3 .955, Hits@1 .949) and FB15k (MRR .795, Hits@10 .892, Hits@3 .833, Hits@1 .741), apart from Hits@10 on WN18 where R-GCN does better.
- TuckER is fully expressive: Theorem 1 gives a sufficient bound of entity embedding dimensionality \(d_e = n_e\) and relation embedding dimensionality \(d_r = n_r\).
- RESCAL, DistMult, ComplEx, and SimplE can be viewed as special cases of TuckER.
- Parameter sharing through the core tensor helps: on WN18RR TuckER uses ∼9.4M parameters versus ∼16.4M for ComplEx and SimplE at \(d_e=200, d_r=30\); at embedding size 20 (∼300k parameters), TuckER is almost as good as ComplEx and SimplE at embedding size 200 (∼6M parameters), and on FB15k it improves Hits@1 by +14% over ComplEx and +8% over SimplE.
Methods
- TuckER factorizes the third-order binary tensor of triples with Tucker decomposition, scoring \(\phi(e_s, r, e_o) = W \times_1 e_s \times_2 w_r \times_3 e_o\), where \(W\) is a core tensor and subject/object entity embeddings are shared; a logistic sigmoid converts scores to probabilities.
- Training uses reciprocal relations, 1-N scoring, Bernoulli negative log-likelihood, Adam with batch size 128, and random-search hyperparameters; evaluation uses the filtered setting with MRR and Hits@k on FB15k, FB15k-237, WN18, and WN18RR.
- Theoretical analysis proves full expressiveness, derives the dimensionality bound, and shows the relationship between TuckER and prior linear tensor-factorization models.
Limitations
- The paper states future work might include exploring how to incorporate background knowledge on individual relation properties; it does not present a dedicated limitations section.
- Evident from the text: evaluation is limited to four standard link-prediction datasets and MRR/Hits@k metrics, with no broader task or computational-efficiency analysis beyond parameter counts.
- Evident from the text: the full-expressiveness bound \(d_e = n_e, d_r = n_r\) is theoretical and may be much larger than practical embedding sizes; the authors expect smaller dimensionalities to suffice in practice.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
5 of 14 cited sources checked · 3 supported · 0 flagged · accuracy not yet measured
Every source the paper cites for this claim was checked and supports it.
In the paper
“The authors show that convolution is a way of introducing sparsity and parameter tying and that HypER can be understood in terms of tensor factorization up to a non-linearity, thus placing HypER closer to the well established family of factorization models.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 1 Introduction
What the sources say
- Supportedcites Balažević et al., 2019 source
Section 4.2 (p. 6) describes HypER as tensor multiplication with a constrained core tensor and notes that ReLU is applied before the final dot product. The Related Work section (p. 3) explicitly identifies that ReLU non-linearity as the key difference from existing factorization approaches and describes convolution as multiplication by a sparse tensor with tied weights. (LLM, conf=0.94)
Strength:boosted· not yet calibratedCausal:correlational· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passes- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“They further prove other models are not fully expressive: DistMult, because it cannot model asymmetric relations; and transitive models such as TransE (Bordes et al., 2013) and its variants FTransE (Feng et al., 2016) and STransE (Nguyen et al., 2016), because of certain contradictions that they impose between different relation types.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 5.1 Full Expressiveness and Embedding Dimensionality
What the sources say
- Not matchedcites Bordes et al., 2013
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not matchedcites Feng et al., 2016
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Supportedcites Nguyen et al., 2016 source
The citation immediately following “STransE” identifies the model, rather than attributing the citing authors’ expressiveness proof to Nguyen et al. The source’s model description (SE section; page unavailable) explicitly introduces STransE. It does not establish the claimed contradictions, but that proof is framed as the citing authors’ own. (LLM, conf=0.91)
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Not matched
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“To our knowledge, ComplEx-N3 (Lacroix et al., 2018) is the only other linear link prediction model that benefits from multitask learning.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.3 Link Prediction Results
What the sources say
- Partly supportedcites Lacroix et al., 2018 sourceExperimental — accuracy not yet measured
The Experimental Setup (p. 7) identifies ComplEx-N3, and the Introduction (pp. 1–2) reports performance gains for ComplEx from modeling inverse predicates and, on harder datasets, from N3 regularization. Our copy does not establish that ComplEx-N3 benefits from multitask learning or that it is the only other linear link-prediction model to do so. (LLM, conf=0.78)
Strength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passes- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“There, rank regularization of the embedding matrices is used to encourage a lowrank factorization, thus forcing parameter sharing between relations.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.3 Link Prediction Results
What the sources say
- Partly supportedcites Lacroix et al., 2018 sourceExperimental — accuracy not yet measured
Section 4 (p. 4) confirms that the paper proposes nuclear 3-norm regularization for tensor-factorization models, including ComplEx. But it does not say that this regularizer forces parameter sharing between relations; Section 2.2 (p. 2) instead describes ComplEx’s parameter sharing as occurring between the subject and object modes. (LLM, conf=0.84)
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passes- Partly supported
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“TuckER outperforms previous state-of-the-art models across standard link prediction datasets, acting as a strong baseline for more elaborate models.”
- Also stated in: § 1 Introduction · plain
“proposing TuckER, a new linear model for link prediction on knowledge graphs, that is simple, expressive and achieves state-of-theart results across all standard datasets”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > Abstract
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- The authors state this as an assumption: Subject and object entity embedding matrices are equivalent, i.e. no distinction is made between the embeddings of an entity depending on whether it appears as a subject or as an object.
- Also stated in: § 1 Introduction · plain
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This is supported by the fact that the margin by which TuckER outperforms other linear models is notably increased on datasets with a large number of relations.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.3 Link Prediction Results
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesThe paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Even with fewer parameters than ComplEx and SimplE at d e = 200 and d r = 30 on WN18RR (∼9.4 vs ∼16.4 million), TuckER consistently obtains better results than any of those models.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.3 Link Prediction Results
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesA result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We can see that the difference between the MRRs of ComplEx, SimplE and TuckER is approximately constant for embedding sizes 100 and 200.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.4 Influence of Parameter Sharing
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesA result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, running their model with equivalent parameter number to TuckER shows comparable performance, supporting our belief that the two models both attain the benefits of multi-task learning, although by different means.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.3 Link Prediction Results
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesA result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is important to note that at embedding dimensionalities 20, 50 and 100, TuckER has fewer parameters than Com-plEx and SimplE (e.g.ComplEx and SimplE have ∼3 million and TuckER has ∼2.5 million parameters for embedding dimensionality 100).”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.4 Influence of Parameter Sharing
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes
- Probably not a substantial claim. 9
- It may state two separate claims as one. 9
Every source the paper cites for this claim was checked and supports it.
In the paper
“As shown in (Trouillon et al., 2017), ComplEx is fully expressive with the embedding dimensionality bound d e = d r = n e • n r .”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 5.1 Full Expressiveness and Embedding Dimensionality
What the sources say
- Supportedcites Trouillon et al., 2017 source
Theorem 4 in §3.2 (p. 14) shows that ComplEx can represent arbitrary families of relation matrices using a shared entity embedding matrix and diagonal relation matrices of dimension nm, where n is the number of entities and m the number of relations. It states that the decomposition rank is at most nm, matching the claimed dimensionality bound for entity and relation embeddings. (LLM, conf=0.92)
Strength:boosted· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, for lower embedding sizes, the dif-ference between MRRs increases by 0.7% for embedding size 50 and by 4.2% for embedding size 20 for ComplEx and by 3% for embedding size 50 and by 9.9% for embedding size 20 for Sim-plE.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.4 Influence of Parameter Sharing
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Results achieved by TuckER are not only better than those of other linear models, such as DistMult, ComplEx and SimplE, but also better than the results of many more complex deep neural network and reinforcement learning architectures, e.g.R-GCN, MINERVA, ConvE and HypER, demonstrating the expressive power of linear models and supporting our claim that simple linear models should serve as a”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.3 Link Prediction Results
Load-bearingStrength:boosted· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Overall, TuckER outperforms previous state-of-the-art models on all metrics across all datasets (apart from hits@10 on WN18 where a non-linear model, R-GCN, does better).”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.3 Link Prediction Results
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“For example, improvement on FB15k is +14% over ComplEx and +8% over SimplE on the toughest hits@1 metric.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.3 Link Prediction Results
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is trivial to show that both RESCAL and DistMult belong to the family of bilinear models.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 5.2 Relation to Previous Linear Models
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“It is interesting to note that the TuckER interpretation of the DistMult scoring function, given that matrices A and C are identical, can alternatively be interpreted as a special case of CP decomposition (Hitchcock, 1927), since Tucker decomposition with a superdiagonal core tensor is equivalent to CP decomposition.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 5.2 Relation to Previous Linear Models
What the sources say
- Not checked — no full text of the sourcecites Hitchcock, 1927 source
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We show that TuckER is a fully expressive model, derive sufficient bounds on its embedding dimensionalities and demonstrate that several previously introduced linear models can be viewed as special cases of TuckER.”
- Also stated in: § 1 Introduction · plain
“Finally, we show that several previous stateof-the-art linear models, RESCAL (Nickel et al., 2011), DistMult (Yang et al., 2015), ComplEx (Trouillon et al., 2016) and SimplE (Kazemi and Poole, 2018), are special cases of TuckER.”
- Also stated in: § 1 Introduction · plain
“showing how TuckER subsumes several previously proposed tensor factorization approaches to link prediction.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > Abstract
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
- The authors state this as an assumption: Subject and object entity embedding matrices are equivalent, i.e. no distinction is made between the embeddings of an entity depending on whether it appears as a subject or as an object.
- Also stated in: § 1 Introduction · plain
In the paper
“Similarly to Com-plEx, Kazemi and Poole (2018) show that SimplE is fully expressive with entity and relation embeddings of size d e = d r = min(n e • n r , γ + 1), where γ represents the number of true facts.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 5.1 Full Expressiveness and Embedding Dimensionality
What the sources say
- Not matchedcites Kazemi and Poole (2018)
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“So far, there have been two possible ways in which linear link prediction models introduce asymmetry into factorization of the binary tensor of triples: • distinct (although possibly related) embeddings for subject and object entities and a diagonal matrix (or equivalently a vector) for each relation, as is the case with models such as ComplEx and SimplE; or • equivalent subject and object entity”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 5.3 Representing Asymmetric Relations
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may state two separate claims as one.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“We use evaluation metrics standard across the link prediction literature: mean reciprocal rank (MRR) and hits@k, k ∈ {1, 3, 10}.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.2 Implementation and Experiments
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“TuckER overcomes this by representing relations as vectors w r , which makes the parameter number grow linearly with the number of relations, while still keeping the desirable property of allowing relations to be asymmetric”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 5.3 Representing Asymmetric Relations
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
- Rests on an assumption: The full rank matrix approach for relations often leads to overfitting, especially for relations with a small number of training triples.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“To test this, we re-implement ComplEx and SimplE with reciprocal relations, 1-N scoring, batch normalization and dropout for fair comparison, perform random search to choose best hyper-parameters”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.4 Influence of Parameter Sharing
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“TuckER achieves state-of-the-art results on standard link prediction datasets, in part due to its ability to perform multi-task learning across relations.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 7 Conclusion
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
In the paper
“We use the filtered setting (Bordes et al., 2013), i.e. all known true triples are removed from the candidate set except for the current test triple.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.2 Implementation and Experiments
What the sources say
- Not matchedcites Bordes et al., 2013
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
- Not matched
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“We choose all hyper-parameters by random search based on validation set performance.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.2 Implementation and Experiments
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
The paper relies on this without support.
Knowledge graphs typically contain only a small subset of all possible facts.
“However, they typically contain only a small subset of all possible facts.”
“However, far from all available information is currently stored in existing knowledge graphs”
The paper relies on this without support.
Manually adding new information to knowledge graphs is costly.
“and manually adding new information is costly”
The authors state this as an assumption.
Subject and object entity embedding matrices are equivalent, i.e. no distinction is made between the embeddings of an entity depending on whether it appears as a subject or as an object.
“Subject and object entity embedding matrices are assumed equivalent, i.e. we make no distinction between the embeddings of an entity depending on whether it appears as a subject or as an object in a particular triple.”
The paper relies on this without support.
The full rank matrix approach for relations often leads to overfitting, especially for relations with a small number of training triples.
“which often leads to overfitting, especially for relations with a small number of training triples.”
The authors state this as an assumption.
TuckER should need a lower number of parameters for obtaining good results than ComplEx or SimplE.
“The ability of knowledge sharing through the core tensor suggests that TuckER should need a lower number of parameters for obtaining good results than ComplEx or SimplE.”
“which supports our initial assumption.”
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
RESCAL parameter count squares the wrong dimension
MinorTable 1 gives RESCAL a term proportional to the square of the relation embedding dimension, d_r, but the paper describes a d_e × d_e matrix for each relation. That description implies a term proportional to the square of d_e instead.
“RESCAL (Nickel et al., 2011) e s W r e o W r ∈ R de 2 O(n e d e + n r d 2 r )”
“the relation matrices W r ∈ R de×de are the d e × d e slices of the core tensor Z.”
The core tensor has conflicting dimension orders
MinorTable 1 gives the TuckER core tensor dimensions as entity × entity × relation, whereas the model definition gives entity × relation × entity. These orders differ when entity and relation embedding dimensions differ, as they do for WN18 and WN18RR.
“W ∈ R de×de×dr is the core tensor of a Tucker decomposition.”
“w r ∈ R dr the rows of R representing the relation embedding vector and W ∈ R de×dr×de is the core tensor.”
TuckER does not outperform HypER on WN18 hits@3
MinorThe results table reports the same WN18 hits@3 value, .955, for TuckER and HypER. This is a tie, although the text says TuckER outperforms previous models on every metric except WN18 hits@10.
“Overall, TuckER outperforms previous state-of-the-art models on all metrics across all datasets (apart from hits@10 on WN18 where a non-linear model, R-GCN, does better).”
“| HypER (Balažević et al., 2019) | no | .951 | 958 | .955 | .947 | .790 | .885 | .829 | .734 | | SimplE (Kazemi and Poole, 2018) | yes | .942 | .947 | .944 | .939 | .727 | .838 | .773 | .660 | | TuckER (ours) | yes | .953 | .958 | .955 | .949 | .795 | .892 | .833 | .741 |”
Citations
Cites
The publisher hasn’t deposited this paper’s reference list.
Research works printed but not linked: 6
- Antoine Bordes, Nicolas Usunier, Alberto Garcia- Duran, Jason Weston, and Oksana Yakhnenko. 2013. Translating Embeddings for Modeling Multi- relational Data. In Advances in Neural Information Processing Systems.
- Rajarshi Das, Shehzaad Dhuliawala, Manzil Zaheer, Luke Vilnis, Ishan Durugkar, Akshay Krishna- murthy, Alex Smola, and Andrew McCallum. 2018. Go for a Walk and Arrive at the Answer: Reason- ing over Paths in Knowledge Bases Using Rein- forcement Learning. In International Conference on Learning Representations.
- Seyed Mehran Kazemi and David Poole. 2018. Sim- plE Embedding for Link Prediction in Knowledge Graphs. In Advances in Neural Information Pro- cessing Systems.
- Maximilian Nickel, Volker Tresp, and Hans-Peter Kriegel. 2011. A Three-Way Model for Collective Learning on Multi-Relational Data. In International Conference on Machine Learning.
- Adam Paszke, Sam Gross, Soumith Chintala, Gre- gory Chanan, Edward Yang, Zachary DeVito, Zem- ing Lin, Alban Desmaison, Luca Antiga, and Adam Lerer. 2017. Automatic Differentiation in PyTorch. In NIPS-W.
- Bishan Yang, Wen-tau Yih, Xiaodong He, Jianfeng Gao, and Li Deng. 2015. Embedding Entities and Relations for Learning and Inference in Knowledge Bases. In International Conference on Learning Representations.
Cited by
Sorted by citations
Items 101–200 of 699
Knowledge graph embedding by reflection transformation
Zhang, Q., Wang, R., Yang, J., & Xue, L. (2021) · Knowledge-Based Systems · doi:10.1016/j.knosys.2021.107861 · Cited by 26
Rule mining over knowledge graphs via reinforcement learning
Chen, L., Jiang, S., Liu, J., Wang, C., Zhang, S., Xie, C., Liang, J., Xiao, Y., & Song, R. (2022) · Knowledge-Based Systems · doi:10.1016/j.knosys.2022.108371 · Cited by 26
“There are many embedding-based [21, 35] or graph neural network-based [42] methods with high performance but low interpretability.”
Block term decomposition with distinct time granularities for temporal knowledge graph completion
Lai, Y., Chen, C., Zheng, Z., & Zhang, Y. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.117036 · Cited by 24
Knowledge Graph Embedding Based on Graph Neural Network
Liang, S. (2023) · IEEE International Conference on Data Engineering (ICDE) · doi:10.1109/icde55515.2023.00379 · Cited by 26
“Factorization models [9], [10] utilize tensor decomposition to formulate the relationship of triplets.”
MethodMulti-Level Interaction Based Knowledge Graph Completion
Wang, J., Wang, B., Gao, J., Hu, S., Hu, Y., & Yin, B. (2023) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2023.3331121 · Cited by 26
da Silva, E. S., Pedrini, H., & dos Santos, A. L. (2023) · IEEE Transactions on Network and Service Management · doi:10.1109/tnsm.2023.3257993 · Cited by 23
“We extended the experiments, reducing the dataset using about 60,000 edges by applying some common link prediction implementations [74], [75], [76], [77], [78], [79], [80], [81], the results are shown in Tables V.”
MethodAdvancements in Complex Knowledge Graph Question Answering: A Survey
Song, Y., Li, W., Dai, G., & Shang, X. (2023) · Electronics · doi:10.3390/electronics12214395 · Cited by 27
Parallel training of knowledge graph embedding models
Kochsiek, A., & Gemulla, R. (2021) · Proceedings of the VLDB Endowment · doi:10.14778/3494124.3494144 · Cited by 23
“, RESCAL [23], TransE [5], DistMult [33] ComplEx [30], RotatE [28], SimplE [15], RGCN [27], QuatE [35], TuckER [4] and HittER [7]–have been proposed in the literature; a key difference between these models is the form of the scoring function.”
BackgroundTBDRI: block decomposition based on relational interaction for temporal knowledge graph completion
Yu, M., Guo, J., Yu, J. X., Xu, T., Zhao, M., Liu, H., Li, X., & Yu, R. (2022) · Applied Intelligence · doi:10.1007/s10489-022-03601-5 · Cited by 26
“In this paper, we introduce TuckER in detail, because TuckER [25] has a certain enlightening effect on our research.”
MethodLearnable convolutional attention network for knowledge graph completion
Shang, B., Zhao, Y., & Liu, J. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2023.111360 · Cited by 23
Mixed Geometry Message and Trainable Convolutional Attention Network for Knowledge Graph Completion
Shang, B., Zhao, Y., Liu, J., & Wang, D. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i8.28745 · Cited by 24
Dihedron Algebraic Embeddings for Spatio-Temporal Knowledge Graph Completion
Nayyeri, M., Vahdati, S., Khan, M. T., Alam, M. M., Wenige, L., Behrend, A., & Lehmann, J. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-06981-9_15 · Cited by 23
Aggregation or separation? Adaptive embedding message passing for knowledge graph completion
Li, Z., Chen, L., Jian, Y., Wang, H., Zhao, Y., Zhang, M., Xiao, K., Zhang, Y., Deng, H., & Hou, X. (2024) · Information Sciences · doi:10.1016/j.ins.2024.121639 · Cited by 24
Improving knowledge graph completion via increasing embedding interactions
Li, W., Peng, R., & Li, Z. (2022) · Applied Intelligence · doi:10.1007/s10489-021-02947-6 · Cited by 22
“TuckER [32] is currently the most powerful model in this category, which is built based on the Tucker decomposition.”
MethodMGKsite: Multi-Modal Knowledge-Driven Site Selection via Intra and Inter-Modal Graph Fusion
Liang, K., Meng, L., Li, H., Liu, M., Wang, S., Zhou, S., Liu, X., & He, K. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3521742 · Cited by 21
“…[24], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43] are naturally suitable for it, such as translational models (e.g., TransE [44]), etc.), tensor decomposition models (e.g., DistMult [45], ComplEX [46], and TuckER [47]) and GNN models (e.g., RGCN [48] and COMPGCN [49]).”
MHRN: A multi-perspective hierarchical relation network for knowledge graph embedding
Xue, Z., Zhang, Z., Liu, H., Li, Z., Han, S., & Zhang, E. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113040 · Cited by 23
Yang, R., Ye, Q., Cheng, C.-L., Zhang, S., Lan, Y., & Jing, Z. (2022) · Evidence-based Complementary and Alternative Medicine · doi:10.1155/2022/8693937 · Cited by 21
“0ere was a corresponding score for each completed triple [25, 27].”
BackgroundDiriE: Knowledge Graph Embedding with Dirichlet Distribution
Wang, F., Zhang, Z., Sun, L. L., Ye, J., & Yang, Y. T. (2022) · International World Wide Web Conference (WWW) · doi:10.1145/3485447.3512028 · Cited by 23
“Matrix decomposition methods like TuckER [5] model that ϕ1 =W ×1 a ×2 f riend ×3 b and ϕ2 =W ×1 b ×2 f riend ×3 a, where ϕ1 and ϕ2 are score functions.”
MethodInfluentialImproving temporal knowledge graph embedding using tensor factorization
He, P., Zhou, G., Zhang, M., Wei, J., & Chen, J. (2022) · Applied Intelligence · doi:10.1007/s10489-021-03149-w · Cited by 21
“ConT [25] can be seen as an extension of Tucker [2] which replaces the core tensor in Tucker decomposition [31] with a time-specific tensor.”
BackgroundMethodMultiview feature augmented neural network for knowledge graph embedding
Jiang, D., Wang, R., Xue, L., & Yang, J. (2022) · Knowledge-Based Systems · doi:10.1016/j.knosys.2022.109721 · Cited by 21
TDN: Triplet Distributor Network for Knowledge Graph Completion
Wang, J., Wang, B., Gao, J., Li, X., Hu, Y., & Yin, B. (2023) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2023.3272568 · Cited by 24
“(cid:2) Conv3D [32]: Conv3D is an extension of ConvE, which employs 3D convolutions to capture deeper feature interactions in KG. (cid:2) TuckER [11]: TuckER is a straightforward and powerful bilinear method based on the Tucker decomposition.”
BackgroundMethodInfluentialIntegrating Entity Attributes for Error-Aware Knowledge Graph Embedding
Zhang, Q., Dong, J., Tan, Q., & Huang, X. X. (2023) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2023.3310149 · Cited by 22
“TuckER [43] is a linear model that offers a relatively simple approach to link prediction in knowledge graphs.”
BackgroundActive knowledge graph completion
Omran, P. G., Taylor, K. L., Méndez, S. J. R., & Armin Haller, A. (2022) · Information Sciences · doi:10.1016/j.ins.2022.05.027 · Cited by 21
Graph Attention Networks With Local Structure Awareness for Knowledge Graph Completion
Ji, K., Hui, B., & Luo, G. (2020) · IEEE Access · doi:10.1109/access.2020.3044343 · Cited by 21
“TuckER [12] learns embedding by outputting a core tensor and embedding vectors of entities and relations.”
BackgroundXu, W., Liu, B., Peng, M., Jiang, Z., Jiang, Z., Liu, K., Liu, K., & Peng, M. (2025) · Information Processing & Management · doi:10.1016/j.ipm.2024.104047 · Cited by 21
GLANet: temporal knowledge graph completion based on global and local information-aware network
Wang, J., Lin, X., Huang, H., Ke, X., Wu, R., You, C., & Guo, K. (2023) · Applied Intelligence · doi:10.1007/s10489-023-04481-z · Cited by 20
Ding, Z., Wu, J., Li, Z., Ma, Y., & Tresp, V. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-43418-1_33 · Cited by 20
Veena, G., Gupta, D., & Kanjirangat, V. (2023) · IEEE Access · doi:10.1109/access.2023.3339552 · Cited by 21
“There are models proposed to identify such relations RotatE [86], MuRE [93], Tucker [94], and PairRE [95].”
BackgroundMethodText-enhanced knowledge graph representation learning with local structure
Li, Z., Jian, Y., Xue, Z., Zheng, Y., Zhang, M., Zhang, Y., Hou, X., & Wang, X. (2024) · Information Processing & Management · doi:10.1016/j.ipm.2024.103797 · Cited by 22
Li, H., Li, Z., Li, K., Rellermeyer, J. S., Chen, L. Y., & Li, K. (2020) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2020.3047460 · Cited by 20
“Sparse Tucker Decomposition (STD) is one of themost popular TFmodels for HOHDST, which can find theN-coordinate systems and those systems are tangled by a core tensor between each other [23].”
BackgroundGuo, J., Zhao, M., Yu, J. X., Yu, R., Song, J., Wang, Q., Xu, L., & Yu, M. (2024) · Information Sciences · doi:10.1016/j.ins.2024.121409 · Cited by 19
Knowledge graph based reasoning in medical image analysis: A scoping review
Huang, Q., & Li, G. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.109100 · Cited by 19
Li, F., Li, X., Wen, S., & Bao, J. (2025) · Advanced Engineering Informatics · doi:10.1016/j.aei.2025.103611 · Cited by 19
Efficient Relation-aware Scoring Function Search for Knowledge Graph Embedding
Di, S., Yao, Q., Zhang, Y., & Chen, L. (2021) · IEEE International Conference on Data Engineering (ICDE) · doi:10.1109/icde51399.2021.00100 · Cited by 22
“…ˆ ˆ O p d q O p N e d ` N r d q Hand-designed DistMult [15] ˆ ˆ ˆ O p d q O p N e d ` N r d q NTN [11] ‘ ˆ ˆ O p d 2 q O p N e d ` N r d 2 q TuckER [20] ˆ ˆ O p d 3 q O p d 3 ` N e d ` N r d q ComplEx [16] ˆ ˆ O p d q O p N e d ` N r d q HypER [25] ˆ ˆ O p d 2 q O p N e d ` N r d q Searched AutoSF…”
BackgroundMethodInfluentialSwift and Sure: Hardness-aware Contrastive Learning for Low-dimensional Knowledge Graph Embeddings
Wang, K., Liu, Y., & Sheng, Q. Z. (2022) · International World Wide Web Conference (WWW) · doi:10.1145/3485447.3511927 · Cited by 19
“Dozens of KGE models have been proposed, which can be divided into three categories from the perspective of the scoring function: (1) geometric distance based models, including TransE [3], TransD [9], RotatE [22], QuatE [35]; (2) tensor factorization based models, including RESCAL [13], DistMult [34], ComplEx [25], Tu…”
MethodTINKER: A framework for Open source Cyberthreat Intelligence
Rastogi, N., Dutta, S., Gittens, A., Zaki, M. J., & Aggarwal, C. C. (2022) · IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom) · doi:10.1109/trustcom56396.2022.00225 · Cited by 36
QDN: A Quadruplet Distributor Network for Temporal Knowledge Graph Completion
Wang, J., Wang, B., Gao, J., Li, X., Hu, Y., & Yin, B. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3274230 · Cited by 23
“The semantic-matching-based methods use the similarity-based score function to evaluate the probability of the candidate triplet, including ComplEx [20] and TuckER [21].”
BackgroundMethodWang, X., Yang, K., Jia, T., Gu, F., Wang, C., Xu, K., Shu, Z., Xia, J., Zhu, Q., & Zhou, X. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae161 · Cited by 21
“Among these typical KGE models, methods based on tensor decomposition perform better (e.g. [48–50], CP-N3).”
Liu, Y., Tian, X., Sun, Z., & Hu, W. P. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-77844-5_11 · Cited by 27
A Knowledge Selective Adversarial Network for Link Prediction in Knowledge Graph
Hu, K., Liu, H., & Hao, T. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-32233-5_14 · Cited by 19
“TuckER [20] is based on Tucker decomposition and the semantic matching models mentioned above such as RESCAL, DistMult, ComplEx and SimplE are all special cases of TuckER.”
MethodBinarized Embeddings for Fast, Space-Efficient Knowledge Graph Completion
Hayashi, K., Kishimoto, K., & Shimbo, M. (2021) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2021.3075070 · Cited by 18
“*, **, *** and **** indicate results transcribed from [10], [24], [25] and [29], respectively. where XX ð b Þ 2 fþ 1 ; (cid:5) 1 g D 1 (cid:4) D 2 is a binary matrix and a is a positive real value.”
BackgroundMADE: Multicurvature Adaptive Embedding for Temporal Knowledge Graph Completion
Wang, J., Wang, B., Gao, J., Pan, S., Liu, T., Yin, B., & Gao, W. (2024) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2024.3392957 · Cited by 19
“McRL [16] captures multiple concepts of each entity/relation hidden in the KG. TuckER [20] introduces the Tucker decomposition into the SKGC methods.”
MethodSelectE: Multi-scale adaptive selection network for knowledge graph representation learning
Zu, L., Lin, L., Fu, S., Guo, F., & Wu, J. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111554 · Cited by 18
Trustworthy Knowledge Graph Completion Based on Multi-sourced Noisy Data
Huang, J., Zhao, Y., Hu, W. P., Ning, Z., Chen, Q., Qiu, X., Huo, C., & Ren, W. (2022) · International World Wide Web Conference (WWW) · doi:10.1145/3485447.3511938 · Cited by 17
“However, the majority of existing methods, such as [1, 3, 5, 8, 10, 11, 16, 25, 32, 34, 39, 42, 47], only address relational fact completion.”
MethodTranslation-Based Embeddings with Octonion for Knowledge Graph Completion
Yu, M., Bai, C., Yu, J. X., Zhao, M., Xu, T., Liu, H., Li, X., & Yu, R. (2022) · Applied Sciences · doi:10.3390/app12083935 · Cited by 19
Each Snapshot to Each Space: Space Adaptation for Temporal Knowledge Graph Completion
Li, Y., Zhang, X., Zhang, B., & Ren, H. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19433-7_15 · Cited by 18
A Simple But Powerful Graph Encoder for Temporal Knowledge Graph Completion
Ding, Z., Ma, Y., He, B., Wu, J., Han, Z., & Tresp, V. (2024) · Lecture notes in networks and systems · doi:10.1007/978-3-031-47715-7_49 · Cited by 18
SANe: Space adaptation network for temporal knowledge graph completion
Li, Y., Zhang, X., Zhang, B., Huang, F., Chen, X., Lu, M. X., & Ma, S. (2024) · Information Sciences · doi:10.1016/j.ins.2024.120430 · Cited by 17
Medical knowledge graph completion via fusion of entity description and type information
Wang, X., Zhang, R., Zhao, B., Yao, Y., Zhao, H., & Zhu, X. (2024) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2024.102848 · Cited by 17
Hou, W., Zhao, W., Jia, N., & Liu, X. (2024) · Applied Soft Computing · doi:10.1016/j.asoc.2024.112622 · Cited by 17
HYPER2: Hyperbolic embedding for hyper-relational link prediction
Yan, S., Zhang, Z., Sun, X., Xu, G., Jin, L., & Li, S. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.04.026 · Cited by 17
MOOMIN: Deep Molecular Omics Network for Anti-Cancer Drug Combination Therapy
Rózemberczki, B., Gogleva, A. A., Nilsson, S., Edwards, G., Nikolov, A., & Papa, E. (2022) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3511808.3557146 · Cited by 16
PosKHG: A Position-Aware Knowledge Hypergraph Model for Link Prediction
Chen, Z., Wang, X., Wang, C., & Li, Z. (2023) · Data Science and Engineering · doi:10.1007/s41019-023-00214-x · Cited by 16
“As for link prediction in knowledge graphs, we compared PosKHG with several baselines, including TransE [3], DistMult [24], ComplEx [23], Sim-plE [11], RotatE [21], TuckER [2], HAKE [31], DualE [5], AutoSF [34], and ComplExRP [6].”
BackgroundMethodInfluentialKnowledge graph representation learning with relation-guided aggregation and interaction
Shang, B., Zhao, Y., & Liu, J. (2024) · Information Processing & Management · doi:10.1016/j.ipm.2024.103752 · Cited by 16
He, J., Liu, J., Wang, L., Li, X., & Xu, X. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10687798 · Cited by 22
HSMH: A Hierarchical Sequence Multi-Hop Reasoning Model With Reinforcement Learning
Wang, D., Li, B., Song, B., Chen, C., & Yu, F. R. (2023) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2023.3303617 · Cited by 17
RoCS: Knowledge Graph Embedding Based on Joint Cosine Similarity
Wang, L., Luo, J., Deng, S., & Guo, X. (2023) · Electronics · doi:10.3390/electronics13010147 · Cited by 17
Distributed representations of entities in open-world knowledge graphs
Guo, L., Chen, Z., Chen, J., Zhang, Y., Sun, Z., Bo, Z., Fang, Y., Liu, X., Chen, H., & Zhang, W. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111582 · Cited by 17
Research Progresses and Applications of Knowledge Graph Embedding Technique in Chemistry
Wang, C., Yang, Y., Song, J., & Nan, X. (2024) · Journal of Chemical Information and Modeling · doi:10.1021/acs.jcim.4c00791 · Cited by 17
Path Ranking with Attention to Type Hierarchies
Liu, W., Daruna, A., Kira, Z., & Chernova, S. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i03.5680 · Cited by 15
KE-X: Towards subgraph explanations of knowledge graph embedding based on knowledge information gain
Zhao, D., Wan, G., Zhan, Y., Wang, Z., Ding, L., Zheng, Z., & Du, B. (2023) · Knowledge-Based Systems · doi:10.1016/j.knosys.2023.110772 · Cited by 17
Is multi-level data enhancement helpful for knowledge graph? A new perspective on multimodal fusion
Yang, K., Yu, R., Guo, B., & Zhen, S. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112285 · Cited by 14
ConvD: Attention Enhanced Dynamic Convolutional Embeddings for Knowledge Graph Completion
Guo, W., Li, Z., Wang, X., Chen, Z., Zhao, J., Li, J., & Yuan, Y. (2025) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2025.3582243 · Cited by 21
“The latest work, TuckER [22], employs a three-dimensional tensor decomposition method to learn vector representations of entities and relations, achieving promising results.”
Imbalanced low-rank tensor completion via latent matrix factorization
Qiu, Y., Zhou, G., Zeng, J., et al. (2022) · Neural Networks · doi:10.1016/j.neunet.2022.08.023 · Cited by 14
Chen, B., Shen, H., Cheng, Z., Liu, X., Zhong, T., & Zhou, F. (2025) · Information Processing & Management · doi:10.1016/j.ipm.2025.104472 · Cited by 14
Semantic- and relation-based graph neural network for knowledge graph completion
Li, X., Tian, Y., & Ji, S. (2024) · Applied Intelligence · doi:10.1007/s10489-024-05482-2 · Cited by 21
Knowledge graph completion using topological correlation and multi-perspective independence
Yu, M., Zhang, Q., Yu, J. X., Zhao, M., Li, X., Jin, D., Yang, M., & Yu, R. (2022) · Knowledge-Based Systems · doi:10.1016/j.knosys.2022.110031 · Cited by 13
MADLINK: Attentive multihop and entity descriptions for link prediction in knowledge graphs
Biswas, R., Sack, H., & Alam, M. (2022) · Semantic Web · doi:10.3233/sw-222960 · Cited by 14
“– TuckER [62] model is based on Tucker decomposition of the third-order binary tensor of triples.”
MethodMQuadE: a Unified Model for Knowledge Fact Embedding
Yu, J., Cai, Y., Sun, M., & Li, P. (2021) · International World Wide Web Conference (WWW) · doi:10.1145/3442381.3449879 · Cited by 13
“TuckER subsumes many bi-linear semantic matching methods and can benefit from the multi-task learning through the core tensor.”
MethodInfluentialDing, P., Pan, Y., Wang, Q., & Xu, R. (2022) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2022.104164 · Cited by 13
BDRI: block decomposition based on relational interaction for knowledge graph completion
Yu, M., Guo, J., Yu, J. X., Xu, T., Zhao, M., Liu, H., Li, X., & Yu, R. (2023) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-023-00918-8 · Cited by 12
“Balazevic et al. (2019) have proved that TuckER is also full expressiveness while indicating that embedding dimensions of entities and relations are limited to n e and n r respectively.”
BackgroundMethodInfluentialDynamic relation learning for link prediction in knowledge hypergraphs
Zhou, X., Hui, B., Zeira, I., Wu, H., & Tian, L. (2023) · Applied Intelligence · doi:10.1007/s10489-023-04710-5 · Cited by 12
“TuckER [19] uses the tucker decomposition method to decompose triplets into a core tensor and three factor matrices.”
MethodGong, C., Wei, Z., Wang, R., Zhu, P., Chen, J., Zhang, H., & Miao, D. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.124652 · Cited by 13
Temporal knowledge graph reasoning based on relation graphs and time-guided attention mechanism
Hu, J., Zhu, Y., Teng, F., & Li, T. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112280 · Cited by 15
ER: Equivariance Regularizer for Knowledge Graph Completion
Cao, Z., Xu, Q., Yang, Z., & Huang, Q. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i5.20490 · Cited by 12
A knowledge graph completion model integrating entity description and network structure
Yu, C., Zhang, Z., An, L., & Li, G. (2022) · Aslib Journal of Information Management · doi:10.1108/ajim-01-2022-0031 · Cited by 12
“2 translation-based models [3-6], and the semantic-matching models [7-16], only use the triple structure in the knowledge graph when training entity and relationship representation, while ignoring many rich information from other sources such as entity description, attributes, etc.”
MethodSem@K: Is my knowledge graph embedding model semantic-aware?
Hubert, N., Monnin, P., Brun, A., & Monticolo, D. (2023) · Semantic Web · doi:10.3233/sw-233508 · Cited by 18
Shan, W., Shen, C., Luo, L., & Ding, P. (2023) · iScience · doi:10.1016/j.isci.2023.108020 · Cited by 11
Learning higher-order features for relation prediction in knowledge hypergraph
Wang, P., Chen, J., Wang, Z., et al. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111510 · Cited by 11
HierarT: Multi-hop temporal knowledge graph forecasting with hierarchical reinforcement learning
Luo, X., Zhu, A., Zhang, J., & Shao, J. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112164 · Cited by 11
HoGRN: Explainable Sparse Knowledge Graph Completion via High-Order Graph Reasoning Network
Chen, W., Cao, Y., Feng, F., He, X., & Zhang, Y. (2024) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2024.3422226 · Cited by 16
KGDA: A Knowledge Graph Driven Decomposition Approach for Cellular Traffic Prediction
Gong, J., Li, T., Wang, H., Liu, Y., Wang, X., Wang, Z., Deng, C., Feng, J., Jin, D., & Li, Y. (2024) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3690650 · Cited by 12
“In particular, we utilize Tucker model [2] and assess the plausibility of a triplet ( ℎ,𝑟,𝑡 ) ∈ F : where W ∈ R × 𝑑 × is the key tensor in Tucker decomposition [36], ℎ , 𝒆 𝑡 , 𝒓 ∈ R are embeddings for entities and relation.”
MCCI: A multi-channel collaborative interaction framework for multimodal knowledge graph completion
Zhang, X., Dang, J., Wang, Y., & Li, S. (2025) · Information Processing & Management · doi:10.1016/j.ipm.2025.104156 · Cited by 12
An Overview of Utilizing Knowledge Bases in Neural Networks for Question Answering
Kafle, S., de Silva, N., & Dou, D. (2020) · Information Systems Frontiers · doi:10.1007/s10796-020-10035-2 · Cited by 11
“Such insights are also able to explain the performance of tensor factorization methods in KB embeddings such as Tucker factorization (Balazevic et al. 2019b; 2019a).”
BackgroundDaruna, A., Das, D., & Chernova, S. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9982104 · Cited by 11
“We used the TuckER [26] KGE, a recent SoTA KGE model based on the Tucker decomposition, to represent KGE Θ learned from D.”
MethodAdaptive pseudo-Siamese policy network for temporal knowledge prediction
Shao, P., Liu, T., Che, F., Zhang, D., & Tao, J. (2023) · Neural Networks · doi:10.1016/j.neunet.2023.01.004 · Cited by 11
MEGA: Meta-Graph Augmented Pre-Training Model for Knowledge Graph Completion
Wang, Y., Ouyang, X., Guo, D., & Zhu, X. (2023) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3617379 · Cited by 11
“…include conventional translation-based KGC methods (i.e., TransE [2], TransH [53], TransR [20], DKRL [61], TransT [22], DistMult [62], RotatE [34], TuckER [1], TransC [47]), HAKE [64] and current SOTA pre-training KGC methods (i.e., KG-BERT[63], MTL-KGC[19], StAR [39], MEM [4], SimKGC [42]).”
MethodDT4KGR: Decision transformer for fast and effective multi-hop reasoning over knowledge graphs
Xia, Y., Luo, J., Zhou, G., Lan, M., Chen, X., & Chen, J. (2024) · Information Processing & Management · doi:10.1016/j.ipm.2024.103648 · Cited by 11
Uncertain knowledge graph embedding: an effective method combining multi-relation and multi-path
Liu, Q., Zhang, Q., Zhao, F., & Wang, G. (2024) · Frontiers of Computer Science · doi:10.1007/s11704-023-2427-z · Cited by 15
Aggregating neighborhood information for negative sampling for knowledge graph embedding
Liu, H., Hu, K., Wang, F. L., & Hao, T. (2020) · Neural Computing and Applications · doi:10.1007/s00521-020-04940-5 · Cited by 12
“Based on Tucker decomposition, TuckER [31] was also expressive and the models such as RESCAL, Dis-tMult, ComplEx and SimplE could be viewed as special cases of TuckER.”
MethodHRER: A New Bottom-Up Rule Learning for Knowledge Graph Completion
Liang, Z., Yang, J., Liu, H., Huang, K., Cui, L., Qu, L., & Li, X. (2022) · Electronics · doi:10.3390/electronics11060908 · Cited by 11
EARR: Using rules to enhance the embedding of knowledge graph
Li, J., Xiang, J., & Cheng, J. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.120831 · Cited by 11
Sub-Entity Embedding for inductive spatio-temporal knowledge graph completion
Wan, G., Zhou, Z., Zheng, Z., & Du, B. (2023) · Future Generation Computer Systems · doi:10.1016/j.future.2023.05.030 · Cited by 10
A collaborative learning framework for knowledge graph embedding and reasoning
Wang, H., Song, D., Wu, Z., Li, J., Zhou, Y., & Xu, J. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111505 · Cited by 11
M 2 ixKG: Mixing for harder negative samples in knowledge graph
Che, F., & Tao, J. (2024) · Neural Networks · doi:10.1016/j.neunet.2024.106358 · Cited by 11
The gene function prediction challenge: Large language models and knowledge graphs to the rescue
Sunil, R. S., Lim, S. C., Itharajula, M., & Mutwil, M. (2024) · Current Opinion in Plant Biology · doi:10.1016/j.pbi.2024.102665 · Cited by 10
“In contrast, other methods represent the KG as a multi-dimensional adjacency matrix (tensor) that can be factorized and decomposed into low-dimensional vectors corresponding to its entities and relationships [59].”
Advancing rule learning in knowledge graphs with structure-aware graph transformer
Xu, K., Chen, M., Feng, Y., & Dong, Z. (2024) · Information Processing & Management · doi:10.1016/j.ipm.2024.103976 · Cited by 12
HOPLoP: multi-hop link prediction over knowledge graph embeddings
Ranganathan, V., & Barbosa, D. (2021) · World Wide Web · doi:10.1007/s11280-021-00972-6 · Cited by 10
“ComplEx [66] and TuckER [4] are derived from tensor factorization methods (ComplEx: [48, 73], TuckER: [67]) which looks to decompose the binary tensor representation of the KG.”
MethodInfluentialDynamic Generation of Knowledge Graph Supporting STEAM Learning Theme Design
Ke, Q., & Lin, J. (2022) · Applied Sciences · doi:10.3390/app122111001 · Cited by 10
IMF: Interpretable Multi-Hop Forecasting on Temporal Knowledge Graphs
Du, Z., Qu, L., Liang, Z., Huang, K., Cui, L., & Gao, Z. (2023) · Entropy · doi:10.3390/e25040666 · Cited by 10
“into geometric models [1,17,20–22], tensor decomposition models [15,16,18,23], and deep learning models [24–26].”
Background
Metadata
- Authors
- Balazevic, I., Allen, C., & Hospedales, T.
- Affiliations
- University of Edinburgh (GB)
- Venue
- Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)
- Published
- 2019
- Pages
- 5184-5193
- OpenAlex
- W2914592219
- Cited by
- 696
- References
- 0
- Citations by year
- 2026: 81 · 2025: 125 · 2024: 165 · 2023: 137 · 2022: 109 · 2021: 55 · 2020: 21 · 2019: 3
- FWCI
- 17.97
- Type
- conference paper
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- Locations
- arXiv (Cornell University) (submittedVersion)
- Edinburgh Research Explorer (University of Edinburgh) (submittedVersion, cc-by)
- arXiv (Cornell University)
- Funders
- University of Edinburgh
- Engineering and Physical Sciences Research Council (EP/L016427/1)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Advanced Graph Neural Networks
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Mathematics › Computational Mathematics › Tensor decomposition and applications
- Keywords
- knowledge graph completion, tensor decomposition, Tucker, knowledge graphs
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-23, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-23 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
