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
22 Own interpretation · 9 Own result · 6 Sources not found · 3 Partly supported · 3 Own method · 3 No basis given · 2 Supported · 1 No full text of the sources · 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)
boostedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- 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)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- 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)
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- 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)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- 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
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“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
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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, 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- 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)
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It 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
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
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
“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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It 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.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It 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.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably 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.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ TuckER: Tensor Factorization for Knowledge Graph Completion > Model
“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 )”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 5.2 Relation to Previous Linear Models
“the relation matrices W r ∈ R de×de are the d e × d e slices of the core tensor Z.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ TuckER: Tensor Factorization for Knowledge Graph Completion > Model
“W ∈ R de×de×dr is the core tensor of a Tucker decomposition.”
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 4 Tucker Decomposition for Link Prediction
“w r ∈ R dr the rows of R representing the relation embedding vector and W ∈ R de×dr×de is the core tensor.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ TuckER: Tensor Factorization for Knowledge Graph Completion > 6.3 Link Prediction Results
“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 > Additional information
“| 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 401–500 of 699
Modifier Embedding for Temporal Knowledge Graph Embedding
Wang, R., Li, T., Luo, F., & Fu, R. (2023) · International Joint Conference on Information and Communication Engineering (JCICE) · doi:10.1109/jcice59059.2023.00040 · Cited by 2
Shi, J., Li, M., & Jing, Z. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-44216-2_35 · Cited by 2
Spatial Link Prediction with Spatial and Semantic Embeddings
Mann, G., Dsouza, A., Yu, R., et al. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-47240-4_10 · Cited by 2
Pang, J., Qin, H., Liu, Y., & Liu, X.-Q. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-46664-9_6 · Cited by 2
Enhanced Multi-Modal Knowledge Graphs Representation Learning via Self-Attention Fusion
Jiahong, T., Yongchuan, Z., & He, Y. (2023) · International Conference on Software Engineering and Computer Science (CSECS) · doi:10.1109/csecs60003.2023.10428415 · Cited by 2
“TuckER [22] employs a third-order tensor decomposition.”
Two-Stage Knowledge Graph Completion Based on Semantic Features and High-Order Structural Features
Ying, X., Luo, S., Yu, M., Zhao, M., Yu, J. X., Guo, J., & Li, X. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-2242-6_12 · Cited by 2
cuFastTucker: A Novel Sparse FastTucker Decomposition For HHLST on Multi-GPUs
Li, Z., Hu, Y., Li, M., Yang, W., & Li, K. (2024) · ACM Transactions on Parallel Computing · doi:10.1145/3661450 · Cited by 2
“It has become a popular technique due to its effectiveness in reducing the dimensionality of tensors [20].”
MethodRuMER-RL: A hybrid framework for sparse knowledge graph explainable reasoning
Zeng, Z., Cheng, Q., Si, Y., & Liu, Z. (2024) · Information Sciences · doi:10.1016/j.ins.2024.121144 · Cited by 2
Zhao, E., Chen, B., BianBaDroMa, ‘., & Ngodrup (2024) · Knowledge and Information Systems · doi:10.1007/s10115-024-02174-8 · Cited by 2
“…textual descriptions, such as DKRL [37], jointly [38], and AATE_E [2], and some recent non-translation models such as meta-TransE [7], TuckER [1, 3], GETD [20], McRL[30] and CKRL[24]. meta-TransE [7] is the metagraph of a knowledge graph by proposing a new affinity metric that measures the…”
Graph Attention Network with Relational Dynamic Factual Fusion for Knowledge Graph Completion
Yu, M., Zuo, Y., Zhang, W., Zhao, M., Xu, T., Zhao, Y., Guo, J., & Yu, J. X. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-70359-1_6 · Cited by 2
Temporal Knowledge Graph Reasoning Based on Dynamic Fusion Representation Learning
Chen, H., Zhang, M., & Chen, Z. (2024) · Expert Systems · doi:10.1111/exsy.13758 · Cited by 4
“Common decoders such as DistMult, TuckER, ConvKB and so forth perform poorly when dealing with complex relational patterns.”
Chen, Y., Chen, G., & Li, P. (2024) · Sensors · doi:10.3390/s24227128 · Cited by 2
Research on Knowledge Graph Completion Method Based on Graph Convolutional Neural Networks
Wang, Y., Feng, S., Wang, R., et al. (2024) · International Conference on Intelligent Computing and Signal Processing (ICSP) · doi:10.1109/icsp62122.2024.10743659 · Cited by 2
Tucker Decomposition-Enhanced Dynamic Graph Convolutional Networks for Crowd Flows Prediction
Dai, G., Kong, W., Liu, Y., et al. (2024) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3706116 · Cited by 2
“However, the ∈ R × 𝑉 × introduces 2 parameters, which easily tend to explode as the number of regions increases.”
Li, H., Wei, J., Yu, J., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-981-96-0814-0_27 · Cited by 2
Dual-view embedding for hyper-relational knowledge graphs with hierarchical structure
Liu, S., Xu, L., Liu, Y., & Kolmanič, S. (2025) · Journal of Intelligent Information Systems · doi:10.1007/s10844-024-00916-2 · Cited by 2
SIKGC: Structural Information Prompt Based Knowledge Graph Completion with Large Language Models
Li, W., Ge, J., Feng, W., Zhang, L., Li, L., & Wu, B. (2024) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc54092.2024.10831590 · Cited by 2
“Such as TransE [6], DistMult [7], ComplEx [8], RotatE [9] and TukER [10].”
Efficient Parameterization for Knowledge Graph Embedding Using Hierarchical Attention Network
Chen, Z.-Y., Liu, F.-C., Wang, X. Y., Lee, C., & Lin, C.-S. (2025) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2025.061661 · Cited by 2
Health Misinformation in Social Networks: A Survey of Information Technology Approaches
Papanikou, V., Papadakos, P., Karamanidou, T., Stavropoulos, T. G., Pitoura, E., & Tsaparas, P. (2025) · Future Internet · doi:10.3390/fi17030129 · Cited by 13
Wang, T., & Shen, B. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06459-5 · Cited by 2
“TuckER [27] relies on tensor factorization and employs a Tucker decomposition [20] of a third-order binary tensor representing the knowledge graph.”
Li, Q., Liang, S., Zhang, Y., Ji, C., Chang, Z., & Wang, S. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3754957 · Cited by 4
Multi-Stage Network Embedding for Exploring Heterogeneous Edges
Huang, H., Yu, S., Ye, F., Xie, X., Shi, X., & Jin, H. (2020) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3415157 · Cited by 2
“Generally, the unified proximity matrix of a graph with heterogeneous edges can be treated as a collection of proximity matrices from different views, thus a great deal of efforts [1, 12, 16, 21] have been devoted to modeling heterogeneous edges by tensor factorization.”
BackgroundMethodScientific and Technological Resource Sharing Model Based on Few-Shot Relational Learning
Liu, Y., Гу, Ф., Ji, Y., et al. (2021) · IEICE Transactions on Information and Systems · doi:10.1587/transinf.2020bdp0021 · Cited by 2
A unified Link Prediction architecture applied on a novel heterogenous Knowledge Base
Hilman, D., & Şerban, O. (2022) · Knowledge-Based Systems · doi:10.1016/j.knosys.2022.108228 · Cited by 2
Modeling N-ary relational data as gyro-polygons with learnable gyro-centroid
Yan, S., Zhang, Z., Xu, G., Sun, X., Li, S., & Wang, S. (2022) · Knowledge-Based Systems · doi:10.1016/j.knosys.2022.109164 · Cited by 2
Zulaika, U., Almeida, A., & López–de–Ipiña, D. (2022) · International Conference on Smart and Sustainable Technologies (SpliTech) · doi:10.23919/splitech55088.2022.9854264 · Cited by 2
“Recently, TuckER [11] was proposed as a generalisation of every other model based on the Tucker decomposition Tucker et al.”
MethodOpen Knowledge Graph Link Prediction with Segmented Embedding
Xie, T., Peng, P., Wang, H., & Liu, Y. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9891940 · Cited by 2
“Typical approaches are TransE [4], DistMulti [5], TuckerE [6], and RotatE [7].”
MethodA new relational reflection graph convolutional network for the knowledge representation
Yao, S., & Pi, D. (2023) · Journal of Ambient Intelligence and Humanized Computing · doi:10.1007/s12652-023-04516-w · Cited by 2
“TuckER (Balažević et al. 2019) is a relatively straightforward but powerful linear model based on Tucker decomposition of the binary tensor representation of knowledge graph triples.”
BackgroundKnowledge Hypergraph Reasoning Based on Representation Learning
Li, Z. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-30678-5_66 · Cited by 2
An Evaluation of Link Prediction Approaches in Few-Shot Scenarios
Braken, R., Paulus, A., Pomp, A., & Meisen, T. (2023) · Electronics · doi:10.3390/electronics12102296 · Cited by 3
Neighboring relation enhanced inductive knowledge graph link prediction via meta-learning
Liu, B., Peng, M., Xu, W., & Peng, M. (2023) · World Wide Web · doi:10.1007/s11280-023-01168-w · Cited by 2
“2)factorization-based models including RESCAL [32], DistMult [33], HolE [34], ComplEx [14] and TuckER [35] utilize tensor decomposition to represent each relation as a matrix to model relation reasoning patterns(e.g., symmetric, reflexive).”
MethodKnowledge Graph Embedding Method Based on Entity Metric Learning
He, M., Duan, L., Zhang, B., & Han, S. W. (2023) · International Conference on Computing and Artificial Intelligence (ICCAI) · doi:10.1145/3594315.3594345 · Cited by 2
“Inspired by literature [18], [24], this paper graphs all the features of relations into dr dimensional vector space, and all the features of entities in de dimensional vector space, and changes the direct co-occurrence of entities and relations into co-occurrence relations between features.”
MethodGood Intentions: Adaptive Parameter Management via Intent Signaling
Renz-Wieland, A., Kieslinger, A., Gericke, R., Gemulla, R., Kaoudi, Z., & Markl, V. (2023) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3583780.3614895 · Cited by 1
“Such sparse access is common in natural language processing [13, 33, 39, 40], knowledge graph embeddings [5, 6, 36, 51], some graph neural networks [45, 47, 53], click-through-rate prediction [11, 15, 56, 64], and recommender systems [9, 19, 24].”
BackgroundYan, X., Yao, Y., Shi, W., et al. (2024) · International Journal on Semantic Web and Information Systems · doi:10.4018/ijswis.336919 · Cited by 2
A unified embedding-based relation completion framework for knowledge graph
Zhong, H., Li, W., Zhang, Q., et al. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111468 · Cited by 2
He, T., Liu, M., Cao, Y., et al. (2024) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-023-01001-y · Cited by 2
“Besides, we compare VEM 2 L with embedding learning methods (Bordes et al. 2013; Sun et al. 2018; Balazevic et al. 2019; Vashishth et al. 2020; Shang et al. 2019), path-based model (Lin et al. 2018) and its improved version for sparse KGC (Lv et al. 2020), text-based method KG-BERT (Yao et al.…”
Link Prediction Based on Data Augmentation and Metric Learning Knowledge Graph Embedding
Duan, L., Han, S., Jiang, W., He, M., & Qiao, Y. (2024) · Applied Sciences · doi:10.3390/app14083412 · Cited by 2
CosUKG: A Representation Learning Framework for Uncertain Knowledge Graphs
Shen, Q., & Qu, A. (2024) · Mathematics · doi:10.3390/math12101419 · Cited by 3
Zheng, Z., Zhou, B., Yang, H., Tan, Z., Waaler, A., Kharlamov, E., & Soylu, A. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-60626-7_6 · Cited by 3
Prior Bilinear-Based Models for Knowledge Graph Completion
Li, J., Luo, R., Sun, J., Xiao, J., & Yang, Y. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-70352-2_19 · Cited by 2
MSGE: A Multi-step Gated Model for Knowledge Graph Completion
Tan, C., Yang, K., Dai, X., Huang, S., & Chen, J. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-47426-3_33 · Cited by 2
Jiang, X., He, K., & Chen, Y. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.118936 · Cited by 1
Improving Hyper-relational Knowledge Graph Representation with Multi-grained Encoding
Ma, T., Huang, L., & Xue, H. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-30672-3_51 · Cited by 1
Han, S., Guan, Z., Li, S., et al. (2023) · IEEE International Conference on Software Quality, Reliability and Security (QRS) · doi:10.1109/qrs60937.2023.00077 · Cited by 1
Efficient Knowledge Graph Embeddings via Kernelized Random Projections
Goyal, N., Goel, A., Garg, T., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-58502-9_14 · Cited by 1
Poster Abstract: Generative Modeling of Post-Disaster POI Visits Recovery
Li, H., Zhuo, Y., Wang, C., et al. (2024) · ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN) · doi:10.1109/ipsn61024.2024.00050 · Cited by 1
Knowledge Management in SMEs: Applying Link Prediction for Assisted Decision Making
Lehmann, S., Schließer, J., Schumann, S., Winkler, H., & Jabs, I. (2024) · Lecture notes in networks and systems · doi:10.1007/978-3-031-67159-3_24 · Cited by 3
Public Behavior and Emotion Correlation Mining Driven by Aspect From News Corpus
Wang, X., Chang, Y., Kou, L., Luo, X., & Zhang, H. (2024) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2024.3441011 · Cited by 1
“1) Baseline Model and Assessment Metrics: Six knowledge representation models are selected for comparative experiments, which includes the classical DistMult [37], Com-plEx [38], Analogy [39], SimplE [40] models, and the recently proposed TuckERL [41] and CTKGC [42] knowledge representation models.”
NSE-KGC: A knowledge graph completion method based on neighborhood semantic evidence
Lv, Y., & Bao, H. (2024) · doi:10.1109/icbase63199.2024.10762633 · Cited by 1
“Our baselines are selected from three categories: embeddin-g-based models: TransE[3], RotatE[4], TuckER[12], SE-GNN[9]; text-based models: KG-BERT[7], MTL-KGC[13], StAR[8], SimKGC[10]; and models that combine structure with text: GenKGC[14], CSProm-KG[5].”
Causal Inference KGC: A Multi-Hop Reasoning Framework Over Sparse Knowledge Graph
Shen, C., Du, H., & Li, P. (2024) · International Conference on Behavioural and Social Computing (BESC) · doi:10.1109/besc64747.2024.10780638 · Cited by 1
“• TuckER [28] introduces a retrieve-and-reason process for question answering using large knowledge graphs and text corpora, integrating textual and structured information.”
Chen, B., Li, H., & Shi, Z. (2025) · IEEE Access · doi:10.1109/access.2025.3535143 · Cited by 1
“For example, the ConvE model [32] employs 2D convolution to capture the intrinsic interactions of facts.”
Remote Sensing Image Dehazing via Progressive Frequency Filtering and Adaptive Self-Attention
Xu, L., Xu, Y., Gong, X., et al. (2025) · IEEE Access · doi:10.1109/access.2025.3543397 · Cited by 1
“2, the progressive feature sharing module [50] is employed to learn the high-frequency feature from the multiple scales.”
Maximum Mutual Information Estimation based Graph Attention Network for Knowledge Graph Completion
Zhang, W., Luo, S., Meng, Z., Zhao, M., Xu, T., Yu, J. F., Mei, J., & Yu, M. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889069 · Cited by 1
“3) Baseline Model: We compare the experimental results of MIEGAT with the following representative knowledge graph completion models: TransE [7], DistMult [10], RotatE [11], TuckER [12], ConvE [8], ConvKB [13], R-GCN [5], HRAN [14], InteractE [15], RGHAT [16], KBGAT [4], DRR-GAT [17] and GATH [18].”
Hard or False: Keep the Balance for Negative Sampling in Knowledge Graphs
Che, F., Tao, J., & Dai, Q. (2025) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2025.3550545 · Cited by 1
“The main tendency of the present KGE methods pay more attention to the score functions [15], [16], embedding spaces [17], [18], embedding distributions [19], [20] or leveraging pre-trained language models [21], [22].”
Xu, T., Wang, Y., Zhang, W., Zhao, Y., Yu, J. F., Yu, M., Guo, J., & Zhao, M. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06520-3 · Cited by 1
“…the translation-based models (TransE [18], RotatE [37], PairRE [36]), tensor factorization-based models (DistMult [19], ComplEx [39], TuckER [41], DensE [42]), transformer-based models (GenKGC [47], PatReFormer [25]), CNNs (ConvE [20], ConvKB [44], InteractE [45]), transformer-based models…”
Knowledge Graph Construction and Complementation for Research Projects
Li, T., Lin, M., Wang, W., et al. (2025) · Journal of Systems Engineering and Electronics · doi:10.23919/jsee.2025.000064 · Cited by 1
Raker: A Relation-Aware Knowledge Reasoning Model for Inductive Relation Prediction
Wang, J., Li, W., Shu, Y., Guan, J., Zhang, Y., & Zhou, S. (2025) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3745029 · Cited by 1
“Embedding-based methods, e.g., TransE [2], TransR [20], RoateE [32], Complex [35], ConvE [7], and TuckER [1], encode entities and relations as low-dimensional vectors to learn their semantic and structural information and design certain score functions to evaluate the possibility of the target…”
Graph Self-attention Mechanism for Interpretable Multi-hop Knowledge Graph Link Prediction
Liu, H., Li, D., Zeng, B., Liang, W., & Li, D. (2025) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3737702 · Cited by 1
“…query relation as a translation between head and tail entities to construct a distanced-based scoring function; (2) semantic-matching-based models [1, 38] that employ three-way tensor decomposition to establish the scoring function to evaluate the potential semantics in embedded entities and…”
Chen, X., Zhou, S., Liang, K., et al. (2025) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2025.3591056 · Cited by 1
“Traditional embedding-based methods, such as TransE [8], ComplEx [9], and TuckER [32], struggle to generalize to previously unseen entities.”
Cornell, F., Jin, Y., Jussi Karlgren, C., et al. (2025) · Proceedings - International Conference on Data Engineering · doi:10.1109/icde65448.2025.00127 · Cited by 1
HSAE: Hierarchical structure augment embedding for various knowledge graph completion
Xue, Y., Cai, W., Ma, Y., Senhadji, L., Shu, H., & Wu, J. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.114320 · Cited by 1
Shen, Q., Zhang, H., Qu, A., & Dong, S. (2025) · Complex & Intelligent Systems · doi:10.1007/s40747-025-02064-8 · Cited by 1
“Examples include models such as DistMult [20], ComplEx [21], TuckER [22], etc.”
DMutDE: Dual-View Mutual Distillation Framework for Knowledge Graph Embeddings
Liu, R., Wu, Z., Wu, Y., Cao, Z., Xu, Q., & Huang, Q. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3608503 · Cited by 3
“TuckER [27] proposes a linear model based on Tucker decomposition applied to a binary adjacency matrix.”
Efficient relational context perception for knowledge graph completion
Tu, W., Wan, G., Shang, Z., & Du, B. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06902-7 · Cited by 1
“When the TRP block encoder is removed, the model can be viewed as an original TuckER model [10].”
InfluentialLink Prediction Under Non-targeted Attacks: Do Soft Labels Always Help?
Memariani, A., Röder, M., Sharma, A., Demir, C., & Ngomo, A.-C. N. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-09527-5_6 · Cited by 2
SPARK: Adaptive Low-Rank Knowledge Graph Modeling in Hybrid Geometric Spaces for Recommendation
Wang, B., Xiao, Y., Wang, M., et al. (2025) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3746252.3760999 · Cited by 1
“To explicitly model the multi-way interactions inherent within KG triples ( ℎ,𝑟,𝑡 ) and crucially, to generate more compact and noise-resilient semantic representations, particularly for less frequent long-tail entities, we strategically employ a scoring function inspired by low-rank Tucker decomposition, as proposed…”
EmbER: A Neuro-Symbolic Method for Entity Comparison and Ranking in Text-Rich Knowledge Graphs
Chen, H., Jones, G. J. F., & Brennan, R. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-13109-6_5 · Cited by 1
Zhu, M., Liu, Y., Luo, Z., & Zhu, T. (2026) · npj Digital Medicine · doi:10.1038/s41746-025-02176-y · Cited by 5
Gao, W., Wang, H., & Yang, H. (2026) · Symmetry · doi:10.3390/sym18020328 · Cited by 1
HyperComplEx: Adaptive Multi-Space Knowledge Graph Embeddings
Gajjar, J., Ranaware, K., Subramaniakuppusamy, K., & Gandhi, V. C. (2025) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata66926.2025.11400828 · Cited by 1
“While tensor-based variants like SimplE [19] and TuckER [20] further improved representation capacity, these Euclidean and complex-valued models remain geometrically constrained, lacking the curvature required for hierarchical reasoning.”
Zhang, A., Lu, X., Cui, Q., Huang, Z., Li, B., Zhang, J., & Ma, K. (2026) · Transactions in GIS · doi:10.1111/tgis.70223 · Cited by 1
“MuRP (Balazevic et al. 2019) projects entities into a non-Euclidean space through projection operations, enabling the capture of diverse relations between entities.”
InfluentialTechnology opportunity prediction based on SAO representation learning
Zhang, J., Zhang, X., Shi, J., et al. (2026) · Scientometrics · doi:10.1007/s11192-026-05578-1 · Cited by 1
MORTIS: Towards Multi-Modal and Multi-Scale Federated Knowledge Graph Completion
Zhang, Y., Li, L., Jin, Z., Chen, Z. Y., Guo, L., Zhang, W., & Chen, H. (2026) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2026.3689321 · Cited by 1
Golovin, A. O., Zhukova, N. A., & Delhibabu, R. (2026) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-026-01206-x · Cited by 1
“TuckER (Balažević et al. 2019) further improves performance through tensor factorization.”
DualKG-DC: A Drug-Centric Dual-Layer Knowledge Graph Framework for Drug Combination Prediction
Gao, Z., Perkins, S. W., Parameswaran, S., et al. (2026) · Journal of Medical Systems · doi:10.1007/s10916-026-02416-y · Cited by 1
“The model represents drug-drug-disease interactions in a tensor formulation and learns latent representations through factorization [29, 30 ].”
Symmetry-Aware Schema–Session Decoupled LoRA Adaptation for Continual Knowledge Graph Completion
Dong, S., Dong, Y., Zhang, H., & Feng, P. (2026) · Symmetry · doi:10.3390/sym18060953 · Cited by 1
Target identification and validation
Ashenden, S. K., Kurbatova, N., & Bartosik, A. (2021) · Elsevier eBooks · doi:10.1016/b978-0-12-820045-2.00005-2 · Cited by 1
Yu, D., Xu, G., Jing, Z., & Li, X. (2021) · IEEE International Conference on Big Data Computing Service and Applications (BigDataService) · doi:10.1109/bigdataservice52369.2021.00026 · Cited by 1
Multimodal learning on graphs for disease relation extraction
Lin, Y., Lu, K., Yu, S., Cai, T., & Žitnik, M. (2023) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2023.104415 · Cited by 35
Text-Enhanced and Relational Context Based Hyperbolic Knowledge Graph Embedding
Ying, X., Li, M., Yu, J. X., Zhao, M., Xu, T., Yu, M., Liu, H., & Li, X. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-10983-6_26 · Cited by 1
Wu, D., Jing, Z., Li, M., et al. (2022) · 2022 International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9892443 · Cited by 1
“TuckER [13]: By introducing TuckER tensor decomposition [19], TuckER learns embeddings by outputting a core tensor and embedding vectors of entities and relations. w ∈ R d e × d r × d e is the core tensor obtained by TuckER decomposition, and d e and d r represent the vector dimensions of entities…”
BackgroundMethodInfluentialDAuCNet: deep autoregressive framework for temporal link prediction combining copy mechanism network
Hou, X., Ma, R., Li, Y., & Ma, Z. (2023) · Knowledge and Information Systems · doi:10.1007/s10115-022-01823-0 · Cited by 1
BuB: a builder-booster model for link prediction on knowledge graphs
Soltanshahi, M. A., Teimourpour, B., & Zare, H. (2023) · Applied Network Science · doi:10.1007/s41109-023-00549-4 · Cited by 1
“In this category, the link prediction problem is converted to a multi-class classification problem, and the multi-class classification loss functions are used (Lacroix et al. 2018; Gao et al. 2021a; Dettmers et al. 2018; Balažević et al. 2019).”
MethodA New Model for Tensor Completion: Smooth Convolutional Tensor Factorization
Takayama, H., & Yokota, T. (2023) · IEEE Access · doi:10.1109/access.2023.3291744 · Cited by 1
“Other important low-rank models are tensor decompositions, with CP decomposition and Tucker decomposition [4], [16], [40] as representative models for image completion [1], [43], [46], [48], [51].”
MethodDual Channel Knowledge Graph Embedding with Ontology Guided Data Augmentation
Song, T., Yin, L., Ma, X., & Luo, J. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-40283-8_32 · Cited by 1
Fusion of Human Cognitive Knowledge and Machine Inference for Breast Cancer Detection
Wang, D., Wang, Y., Wang, Y., Liu, L., Li, J., & Huang, Q. (2023) · International Conference on Advanced Robotics and Mechatronics (ICARM) · doi:10.1109/icarm58088.2023.10218759 · Cited by 1
“In the baseline model, we selected classic end-to-end image recognition networks: ResNet-50 [25], VGG-16 [26], and ViT-B/16 [27]; For the verification of the reasoning method based on the knowledge graph, we adopted the TuckER algorithm based on tensor decomposition [28], TransX [29-31] series algorithm based on trans…”
MethodFully Automated Interpretable Breast Ultrasound Assisted Diagnosis System
Wang, D., Wang, Y., Wang, Y., Liu, L., Li, J., & Huang, Q. (2023) · International Conference on Advanced Robotics and Mechatronics (ICARM) · doi:10.1109/icarm58088.2023.10218807 · Cited by 1
“TuckER knowledge tensor decomposition method decomposes a tensor into a group of matrices and a smaller core tensor, and its evaluation function is defined as: orsos eweWere 321),,( ⑴ ed os Ree , is the embedding vector of head entity and tail entity respectively, rdr Rw is the embedding vector of the relation, ere dd…”
MethodCold-Start Multi-hop Reasoning by Hierarchical Guidance and Self-verification
Xu, M., Sun, K., Li, Y., & Qian, T. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-43415-0_34 · Cited by 1
“Knowledge graph embedding methods [3–5, 12, Multi-hop reasoning is an emerging task for KGC, which aims to find the target tail entities with interpretable reasoning paths.”
MethodTKGF-NTP: Temporal Knowledge Graph Forecasting via Neural Temporal Point Process
Han, G., Chen, W., Zhang, X., et al. (2023) · IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws60048.2023.00051 · Cited by 1
Textual Entailment for Effective Triple Validation in Object Prediction
García-Silva, A., Berrío, C., & Gómez-Pérez, J. M. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-47240-4_5 · Cited by 2
“Knowledge Base Completion KBC [3] aims at predicting relations between existing entities.”
BackgroundDynamic Weighted Neural Bellman-Ford Network for Knowledge Graph Reasoning
Liao, H., He, S., Yao, X., et al. (2023) · Communications in computer and information science · doi:10.1007/978-981-99-7224-1_1 · Cited by 1
Temporal Interaction Embedding for Link Prediction in Global News Event Graph
Yang, J., Yang, L. T., Wang, H. B., & Gao, Y. (2024) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2024.3357696 · Cited by 1
“TNTComplEx [21] expands the third-order tensor in ComplEx [27] to fourth-order in the complex space and takes nontemporal predicates into consideration.”
BackgroundMethodLink Prediction Based on Feature Mapping and Bi-Directional Convolution
Feng, P., Zhang, X., Wu, H., Wang, Y., Yang, Z., & Ouyang, D. (2024) · Applied Sciences · doi:10.3390/app14052089 · Cited by 1
Lightweight Multi-Semantic Hierarchy-Aware Poincaré Knowledge Graph Embeddings
Dong, Z., Lin, Y., Tan, H., Wang, L., & Gu, Z. (2023) · IEEE International Conference on Parallel and Distributed Systems (ICPADS) · doi:10.1109/icpads60453.2023.00194 · Cited by 1
“To ensure a fair comparison, we consider a set of benchmark models including TransE [5], DistMult [26], ComplEx [23], ConvE [7], RotatE [22], TuckER [2], HAKE [27], MuRP [1], AttH [6], and HBE [19].”
MethodReliK: A Reliability Measure for Knowledge Graph Embeddings
Egger, M. K., Ma, W., Mottin, D., Karras, P., Bordino, I., Gullo, F., & Anagnostopoulos, A. (2024) · International World Wide Web Conference (WWW) · doi:10.1145/3589334.3645430 · Cited by 4
“• TuckER [4] uses the Tucker decomposition [43] to compute the embeddings for entities and relations.”
MethodLiu, X., Rzepka, R., & Araki, K. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111883 · Cited by 1
Link Prediction in Knowledge Graph with Feature Enhancement
Liang, Y., Zhao, Y., & Yao, Y. (2024) · Mechanisms and machine science · doi:10.1007/978-3-031-70235-8_47 · Cited by 1
Multi-View Few-Shot Reasoning for Emerging Entities in Knowledge Graphs
Yan, C., Zhao, F., Tao, X., & Zhu, X. (2024) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2024.3453749 · Cited by 1
“Previously proposed methods can be broadly classified into translational distance-based [8], factorization-based [16], and neural network based [17] methods.”
Knowledge graph embeddings based on 2d convolution and self-attention mechanisms for link prediction
Zan, S., Ji, W., & Zhou, G. (2024) · Applied Intelligence · doi:10.1007/s10489-024-05977-y · Cited by 1
“TuckER [14] decomposes the tensor corresponding to a triple into a set of vectors and a shared core tensor W .”
InfluentialZhang, H., Han, Q., Sun, H. Z., et al. (2024) · International Conference on Behavioural and Social Computing (BESC) · doi:10.1109/besc64747.2024.10780531 · Cited by 1
“ TuckER [18]:TuckER model is a knowledge graph embedding method based on tensor decomposition.”
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
